As one of the research direction in the field of artificial intelligence, the computer game has a great practical significance. Checkers become a represent of complete knowledge game that because of it has few chess pieces, and moves easily, So it can reflect the thought of the computer game better. First, the recursive algorithm of artificial intelligence can achieve the independent moves and checkmates in draughts game program. Second, the optimization of evaluation function makes the program choose beneficial move ways. Third, the game tree can improve the level of computer in-depth thinking, simplify the search scope, and save computer thinking time by pruning, which improve the level of chess program. In this paper, overall framework of intelligent checkers is built from board representation, game tree, valuation optimization. Based on the accumulation of experience, algorithm upgrade, depth analysis of existing framework,then we give a more optimized system framework and then strengthen the computer game level.