For the collaborative operation of multi-storage robots, we propose a path planning method based on global planning and local adjustment to obtain a short feasible path free of collisions and obstacles. In the path planning, according to the distance from the current node to the end point and the angle between the local path and the Euclidean path from the start point to the end point, we design a new heuristic function which drives the robot to travel along the shortest path. According to the number of optional nodes, we propose the obstacle avoidance rules and improve obstacle avoidance ability. Based on the path length, we compare the pheromone and update it to refine the search space and improve the convergence performance. We also improve the ant colony algorithm to find the optimal path. We design the collision avoidance rules to avoid the collisions between warehouse robots and find the optimal or near-optimal path combination. The experimental results verify the feasibility and effectiveness of the method.