To solve the problems of the slow convergence speed in ant colony algorithm and the local optimum in artificial potential field method, an improved ant colony optimization algorithm was proposed for path planning of mobile robot in the environment expressed by the grid method. The local force factor of artificial potential field was converted into spreading pheromones in the ant searching process, so the ant colony algorithm focused on subspace search with high fitness. It reduced the partial cross paths and the number of lost ants in the process of general ant colony algorithm in blind search. It also enhanced the ability of robot to avoid obstacle in advance. Two algorithms simulation results under different parameter combinations showed that the improved ant colony algorithm not only solved the local optimum problem of artificial potential method, but also avoided the blind search of general ant colony algorithm. In addition, the simulation results were compared with other improved algorithms. The comparisons verified the efficiency of the proposed algorithm which shows better search performance and stronger searching ability than the traditional ant colony algorithms and other improved algorithms. The convergence speed of the proposed algorithm was nearly doubled.