You can't prepare, that's the idea. You get the theme and start from scratch 
But I found it helpful to give myself
- one week for "toying with theme, trying things out",
- then one week of "making a functional playable game"
- and one week of "refining, testing, improving".
That way you have small goals and, most importantly, you have a safety net of seven days should you hit a roadblock.
If you are new to MAGS, you may want to try my approach: Small-scale, single puzzle idea. Then layer stuff on it as long as you have time. A month isn't much time to make a game but is is enough time to make a small, nice-looking game. You got some experience already, so you should know where your strengths lie, and make use of them.
I also tend to really try to get the game done first. No fancy animations and stuff before I can play through a puzzle sequence.
A calendar is useful to have- just mark your "milestone" days red, and cross days out as you make progress.
It's also very, very useful to have a good beta tester, probably one you have worked with before. I remember Chance Of The Dead, where fixing stuff actually took longer than writing the whole game and without a tester who gave detailed feedback and kept suggesting stuff it would've been a really lame game. So get in contact with a good tester.

But I found it helpful to give myself
- one week for "toying with theme, trying things out",
- then one week of "making a functional playable game"
- and one week of "refining, testing, improving".
That way you have small goals and, most importantly, you have a safety net of seven days should you hit a roadblock.
If you are new to MAGS, you may want to try my approach: Small-scale, single puzzle idea. Then layer stuff on it as long as you have time. A month isn't much time to make a game but is is enough time to make a small, nice-looking game. You got some experience already, so you should know where your strengths lie, and make use of them.
I also tend to really try to get the game done first. No fancy animations and stuff before I can play through a puzzle sequence.
A calendar is useful to have- just mark your "milestone" days red, and cross days out as you make progress.
It's also very, very useful to have a good beta tester, probably one you have worked with before. I remember Chance Of The Dead, where fixing stuff actually took longer than writing the whole game and without a tester who gave detailed feedback and kept suggesting stuff it would've been a really lame game. So get in contact with a good tester.
