Abstract: The ant colony algorithm(ACA) has a good global convergence capability by using the mechanism of positive feedback,while genetic algorithm(GA) has a capacity for performing global searches and being quick.CACGA(ant colony-genetic algorithm with adapting parameters based on cloud models) is proposed to take advantage of good qualities of the two optimization algorithms more completely.CBACGA makes the ant colony strategy and the genetic strategy to be fused ingeniously through the cloud association rule,which can utilize the whole function of the algorithm effectively and can dynamically appease the contradiction between the convergent speed and the searching scope.The simulation result for TSP shows its validity.