Yeah, once you have the proposal, go nuts with the code - just don't change the idea, because opne of the criteria will, more than likely, be something along the lines of "Does the project do what it set out to do", so make sure your program fits the description (it can do extra, and this will only gain you marks). No less, though.
But to 94... don't make the project the sole source of your attention. Do other assessment tasks with just as much obsession, make sure you top the class, and go over notes and past exams until you're scoring fantastically on past papers (and no cheating on past papers, I did this for a paper and then realised how dumb it was).
But yeah. Once the project is done, and do it early, do theoretical stuff rather than learning to code - get good at exam technique, e.g. pulling at least 19/20 on the (usually easy) MC questions, and not writing too much for the longer response questions.
If, for example, a question is worth 2 marks, do not raise 3 points. Raise 2 points, and explain them well, in the available space.