Others have already said to keep the project realistic, so I'd like to emphasize a particular point of that to build on as they covered it otherwise rather well:
It's kind of a vague word, but ultimately, it means to keep in mind the total amount of work involved to complete the project down to a level you can accomplish.
Obviously, there's the cost-effort triangle which influences what scope is reasonable. You know the one, "It can be fast, cheap or good - pick two". If you're able to take your time on a project, for example, then you can definitely do stuff in a much wider scope than if you're trying to crunch something out for a game jam.
But all that is just to help you think about the scope of your game and keep it as reasonable as possible. Separate out features that MUST happen(not much of an FPS if you can't shoot things) and WOULD BE NICE to happen(full crafting system allowing you to make fancy cakes) and keep them separate. Same thing with art, sound, etc. You need the player to have a walk cycle, but it doesn't have to be so smooth and flawless as to bring tears to the eye and so on.
Once you have a good breakdown of the needs, you'll usually find that the scope of what you need to do to get a game out is a lot smaller. And from there, you can begin to tackle parts of that list knowing that you don't have to worry about literally everything being done - at least not yet.
Managing scope is absolutely a skill and it's something that takes time and practice to learn. Don't expect to do perfect on your first shot, just understand that if something goes wrong or acts as a roadblock you an use it to learn and be even better the next time something comes up