Some key things to do when creating a game is to spend plenty of time researching and planning how you want your game to look like and function. This probably sounds like "duh" concept, but it's easy to get lost in your game and forget about what you are trying to accomplish with it. As you work on something, you get more ideas and keep changing your game as you go along, and after so many changes, you forget what you were doing in the first place.
You should have a structured folder dedicated to research pertaining to your game. This could be concept art you found on the internet and what to base designs off of, the gameplay or art style of another game you want to mimic, or some fancy game mechanic you want to implement.
Research is the first thing I do when I get an idea of a game. I review everything that I would like to do and save it and document it to look at later. The purpose of this is so I know what I was trying to accomplish and what direction I want to game to go.
After you have done the research, the next step is to plan out the game and everything about it. Take a look at a game that was in the preliminary stages of development.
This is pretty basic, but you get the idea of what you should be looking at. You want to breakdown all of the components of a game and how it will work together. For extra visuals, you can add some of the research content you found earlier to add to this list.
I always start out with a version 0.0, and I use this for the basic setup for starting the game. As I progress through the main stages of building, I add a new line for the version and new number. By doing this you can keep track of all the changes you make for future reference.
It is a good idea to update this even after you release a game in the event you make small adjustments to update your game. You could clean up some code issues, fix some graphics, or anything really. Regardless, having everything planned and typed up will keep you on track with your own deadlines you give yourself.