Abstract: Ant Colony Optimization(ACO) has successfully solved series of discrete optimization problems. However its global conver-gence is not fully studied and proved. This paper proposed an improved Ant Colony Optimization based on cellular automata——Cellular Ant Colony Optimization(CACO).Through redistributing the optimal solution, the algorithm is improved on the speed of convergence. After comparing it with traditional Ant Colony Optimization(ACO), Round Robin under the simulator platform CloudSim, the experiment shows that the algorithm could reduce the whole makespan and decrease the costs of users, which is an effective resource schedule satisfying cloud environment.