I don't understand something about it.

I give the computer a position and let it think for say 120 seconds. It comes up with many black/white moves. Say black is to move. I see what it thinks black should do then whites first reply. So I cancel out of that and make the first black move only. Then I let it think for another 120 seconds for whites move. It comes up with a different move for white than the previous thought process. I don't understand why or how that can be. The whole lines seems different than the first line calculated just one ply ago.

thanks,

I don't see the problem...

If I'm understanding this correctly, it is natural that the second proposed move for White would change... the first proposed move for White was based on Black's position; since you changed that, the second proposed move for White is now based on Black's new position, not the old one.

If CM calculates the "best" moves from white and black, how could the second move change? If black is calculated to move ... Be6 and calculates white to move Nf2, are these the best two moves in CMs estimation? If so, then when I manually do the ... Be6 move, how can it then calculate something other than Nf2 for white the second time around when the first time around Nf2 was the best move?

I just don't understand

When you made one move, Chessmaster can calculate one move deeper, so it may change it's evaluation because of that. It certanly doesn't/shouldn't happen all the time.