In order to find the optimal path for the intelligent inspection robot, the traditional ant colony algorithm is improved .A* is used to improve the ant"s path-finding rules, and multiple smooth iterations are performed on the found path, which improves the defect that the traditional A* algorithm is easy to fall into a "trap". At the same time, the purpose of smoothing the path and shortening the distance is also achieved. Meanwhile, the pheromone processing method of ant colony is optimized, and the double pheromone strategy is introduced to get the optimal path. Compared with the traditional ant colony algorithm, it can avoid obstacles accurately, and has a faster and better ability to search for the optimal path.