Ant colony algorithm (ACO) is a mellow method for planning route for robots, but it has some weak points like slow convergence and poor path optimization ability. The new algorithm proposed a modified method in this paper to overcome these shortcomings. This method has two steps. It uses bird swarm algorithm (BSA) to build the pheromones distribution and then uses ACO to search route. It was found that this method could build a shorter path effectively under the same condition. The algorithm combines ACO, BSA and adaptive expected function, which could enhance the effectiveness of the algorithm. The results showed that in the case of large number and denser arrangement of obstacles, the route in this algorithm is shorter and smoother, and has fewer turning times than ACO, BSA, PSO and the algorithm composed of PSO and ACO. And it has stronger search ability in extreme problems of round about roads and large sunken obstacles.