What I mean is, during development, you will run into a planned feature that proves too difficult to implement, or that doesn't turn out to be fun. You'll be frustrated, because without this feature your game won't match the perfect vision you have in your head. It will be ruined.
The first thing you should do if you start thinking like this, is stop feeling bad about it. Don't start a cycle of depression that could result in you losing all motivation. This has almost happened to me several times while working on my current project, and I don't want it to happen to you. The best thing you can do at that moment, is let it go.
That's right, forget about it for now. Don't stop working on your game, just focus on another piece of it for a while. The creative process doesn't run on a waterfall method. You can't start with a design document for a piece of art, implement it as specified, and release it in a perfect state. Creativity is fickle and ever-changing. Your ideas will shift and warp during the development process, and you should let them. I'm not saying let your mind run wild with possibility and let scope creep take over. I'm just saying to give yourself a chance to work through the possibility of changing your grand vision.
I read some advice in an article once, and some version of it has stuck with me. It's "let the game tell you what it needs". I've interpreted this to mean - play your game, and always be reevaluating. If something is not working, would take too much time, or you don't possess the skill to accomplish it, you need to quickly decide either to give it up or figure it out. It's a tough decision to make, and both options have their merits depending on the situation. You can postpone the decision, but until you make it your brain won't let it go. The most important thing is to make that decision and move on, instead of giving up on the entire project.
I'm starting to ramble a bit so I'll try to wrap up my thoughts. I've found that it helps to go into a project with an understanding that you don't fully understand your idea yet. Give yourself time to prototype and experiment. Brutally cut features when you need to, and always be open to adding features that the game tells you is necessary.