HOW TO NOT FALL INTO THE TRAP OF MAKING PRETTY BUT VAPID GAMES?

Treat the character like a living person. Give them real reactions to things. Give them a personality. Do not fret over the differences between genders or bullshit like that - kids under 10 don't have a defined gender issue unless they've been brought up to. Keep that in mind.

Talk to some kids. Get in their heads.
Like, seriously, get to know the subjects a little. It's something you should do if you're serious about portraying the character correctly.

I'm doing the same kind of thing right now, just without the "pretty" part (unless you love the Rm2k/3 RTP like I do). And, sure, I worry that the depth will be lost on the layman (or ignored completely ("Rm2k3 rtp epic about dragons, yeah")), but I've found that just letting the heavy details and the flow of the story be ultimately helps it feel more natural instead of stilted like other games tend to be. Very rarely have I found that a line of dialogue that I originally wanted needed to be cut because it no longer felt anything but forced in the scene which it appears.

Take a moment to step away from the computer, sit on your bed, and visualize yourself in the scene and talk it out. Take note of where your head is naturally inclined to turn to map out where the characters will be in correlation to each other. This also helps if a character voices a concern as you will instinctively respond to it (facial gestures/head motions speak volumes) as if it were really happening*. You already know where the scene will logically conclude and the ins and outs of your character (if you haven't gotten this far, you need to brainstorm), now you just have to make it flow.

With gameplay, it all boils down to establishing your minimum viable product. Once you've found a method to engage the player simply by playing your game, you can then tack on all the extra bells and whistles that will make your game unique.

* I've actually included pauses or bits of dialogue strictly as a result of a considerate head nod.

In my opinion? Innovate. Create something new, something fresh, unlike the others. I think that's the reason why Dreaming Mary succeeded. It had a very unique atmosphere.

For gameplay, you could couple that up with some new stuff that makes sense. I'm not sure what the context of your game is, but I'd assume it has no combat like the first. (I may be wrong.) But in either case, games like Ib and The Witch's House are good templates for games with art and atmosphere that are also complemented with solid gameplay. Varied, unexpected, and dynamic rather than static.

You can also tell your story through the environment, too. It doesn't have to be very much. Surreal games can thrive off of their ambiguity. Then again, you could break the mould by providing a great amount of depth to a very vague genre by nature (see: OFF).

I may have misinterpreted the question, but in essence, think about the "great game" you want it to be, and then make it. You certainly have the skill.

Gameplay:
I wouldn't try to force your game into any particular suggestion from us. Like, you mentioned adventure games, so the first thing that popped into my mind is designing item puzzles around using and combining items in ways that you usually wouldn't. I half-remember a puzzle from the Curse of Monkey Island that set up the following problem: you have to get an item by a shopkeeper. In order to do this, you have to somehow combine chewing gum, a helium tank, and the item that you want so that you blow a bubble and the item drifts over his head outside his field of vision (again, this wasn't the exact puzzle, but that's the kind of thinking that makes those puzzles rewarding to complete).

However, you don't have to do those kinds of puzzles. Right now, the only person who knows what your game is really about is you. Whatever it is that you want to convey, use the gameplay to help deliver that message. So, if item puzzles don't factor into that, don't tack them on. Try to deliver your message through gameplay--whatever that is. For Yume Nikki, it was just walking around--that was enough.

If you want to work inside an existing genre, then play a lot of games from that genre and ask yourself why you find aspects of it fun. Then, ask yourself how the developers have achieved that. For example, in Super Metroid, the fun might come from finding things. How did the developers make it fun to find things? They hid things everywhere. They gave you lots of tools to find them. They designed the levels to encourage exploration and backtracking. They gave the world an alien feel, so everything feels novel, and so on.

Story:
Writing is difficult. I think it's something that a lot of people jump into with no real experience but think they can because they have the misconception that good writing comes from good ideas. That's not the case. Good writing comes from years of thought and reflection on writing and on people. If you don't understand people and their motivations, you're never going to be a good writer. If you don't take the time to write, study writing, and have your writing critiqued, you're never going to be a good writer.

That said, I'll give some things to remember when developing characters. This'll be a good starting point, and it'll give you some things to think about when you work on your own writing and see the writing of others.

Round/Flat: These are probably terms you're familiar with, but I'll talk about them anyway. A flat character is one that doesn't change over the course of a story. The Terminator from the first film is a good example of this: he's an evil robot at the beginning and end. A round character is a character that changes over the work. The Terminator from the second film is an okay example of this, but we'll stick with him for the sake of comparison: as he interacts with John Connor, he becomes a bit more human over the course of the film. When you write, try to make as many characters as you can ROUND. In a full-scale rpg, there are naturally going to be town npcs that stay flat, which is understandable. Your primary cast, though, needs to be round. Have them change over the course of the game. Look at Walter White at the beginning and end of Breaking Bad. There's a round character. Round characters keep us engaged in the character, and it makes the events of the plot have meaning, as the plot is changing the character.

Make your characters "real." This is what Liberty was getting at. Let reality inform your characters in some way. This doesn't mean that everyone has to be someone you know, but it means they should be human. If they're kids, give them authentic children's voices. If they're teenagers, make them sound like teenagers. Don't fucking Dawson's Creek me. If you have an alien from the planet Zebulon-6, you still give him something human for the audience to connect with. Maybe he feels like an outsider on this planet. Maybe his warlike aggression connects with our own desire to make our own stamp on life. That's kind of something else, but I'll say it here: understand your character's motivations. Everyone wants something out of life: to be desired, to build a legacy, to escape reality, to hide from their own perceived failures. Think about "big picture" motivations--the psychological kind--and use them to inform more surface level things, e.g. a character might want to get rich. That's surface-level. They might want to get rich because they want to escape poverty. That's backstory. The poverty made them feel less than human, so their struggle is built on their desire to change their self-concept. That's the psychology behind it that you need to really understand in order to inform the character.

Intertextuality: This just means that you should understand other works--literature, film, music, and most importantly other games--and how your game fits within those works. The people who pick up your game will automatically compare it to everything they've experienced before, so you should have a good understanding of how your work fits in. If you're smart, you can use and subvert those expectations. This affects writing in general, but it extends to character.

Nuance: This goes with intertextuality, but I figured I'd separate it. Make your characters unique. If your character's a cliche, people will pick up on that. I don't like thinking of writing as "how can I subvert this trope?" I think that's a way you can write and it's not necessarily bad; you should definitely avoid the pitfalls of cliche. However, I think it's better just to really understand your character--to make them human. I think, by doing that, you'll write a nuanced character, anyway, and if you're paying attention to intertextuality, you'll avoid stock characters naturally.

Compact meaning where you can. This is more of a general writing tip, but it applies to characters and their actions. Try to use every moment in your game to function in multiple ways. If an event moves the plot forward, that's great. If it also develops character, then your scenes start to feel charged. If you can compound other things like intertextuality in the mix, then all the better. Well-constructed things show that there's a lot of thought behind the work--that should be something that you already understand from art.

Make everything count. Try to pare down your dialogue. It's really easy to just write a bunch of dialogue that's funny but inherently meaningless. This goes a bit with the last point, but there doesn't necessarily have to be layers of meaning; just make your dialogue important. Interlace meaning in every bit of your work--art, music, story, and gameplay.

Use subtext. Your dialogue should generally avoid having people express their exact thoughts and feelings at all times. Sometimes, this sort of thing is unavoidable and necessary, and it can still be charged dialogue when it's of matters of importance. However, in the bulk of your dialogue, your characters' motivations and feelings should be mostly hidden beneath the surface. It should be implied. Good dialogue makes the reader want to dig into the characters' minds.

Whew, that was a lot. I think those are the main things I think about when developing characters, but there's probably more that I missed. I hope that helps a bit!

Housekeeping, you should really make that an article. That is good stuff.

I was wondering if it is possible to make a Zelda game with VX Ace engine using Lttp resources.
I'm good at designing, eventing, and mapping towns and overworld
but not so good at writing complex stories full of emotional dialogue.