Addressing the problems of deadlock and poor local path in traditional artificial potential field algorithm, some improvement measures were put forward. The obstacle detection algorithm was used to identify one effective obstacle and one intermediate point. Then a local path from starting point to the intermediate point was planed according to the gravitational field and boundary conditions. Setting the intermediate point to a new starting point and repeating this process until each local path was planed one by one. Secondly, aiming at the disadvantage of slow convergence rate and easy to fall into local optimum in basic ant colony algorithm, some improvements were proposed. The result of artificial potential field algorithm was used to build heuristic information of ant colony, so as to avoid the problems of path crossover and slow convergence. At the same time, a negative feedback loop was built to adaptively adjust the renewal speed of global pheromone and local pheromone through the iteration number. Finally, simulation experiment on three different algorithms was conducted. The results showed that under the same environment model, the proposed algorithm had fewer iterations, shorter running time and better global search ability than other two algorithms. In the given simple environment model, the iteration times of the algorithm was 3, the running time was 0.892s, and the optimal path length was 28.627m. In the given complex environment model, the iteration was 8 times, the running time was 3.376s, the optimal path length was 31.556m, and the global coverage of paths was 73.63%.