Abstract: With the continuous advancement of intelligent mine construction in China, the unmanned transportation link has developed into an important part of the intelligent mine system. Scenarios such as the loading and unloading area of open-pit mines are usually unstructured operating areas with complex terrain environment and many obstacles. As the main tool for material transportation of open-pit mines, unmanned mining trucks are more difficult to plan paths in this scenario due to their size, heavy load and other characteristics. In order to solve the problem of low driving efficiency and poor path quality caused by excessive detour during path planning, a "3D-like" path planning method based on optimized ant colony algorithm was proposed, and its effectiveness was verified by simulation and experiment. Firstly, a 3D map construction method based on laser point cloud is designed, and the valid point cloud data after filtering and registration are rasterized and the grid height is calculated, and the 3D map containing obstacle height information is obtained. Secondly, taking unmanned mining truck as the research object, a 3D collision detection method is designed, which can judge the conflict relationship between obstacles and vehicle body in the horizontal and vertical aspects respectively, and according to the structural characteristics of mining truck and road conditions, a parallel crossing strategy is developed to directly cross over obstacles that are not threatening to vehicles, which can effectively improve the passing efficiency of mining truck under the premise of ensuring safety. Then, the initial pheromone distribution of ant colony algorithm is optimized to improve the goal orientation of the algorithm, and the optimal and worst path is considered in the improved pheromone updating strategy to improve the performance and efficiency of path search. Adaptive multi-step movement mode is introduced, and a multi-objective heuristic function is designed to introduce cross-obstacle evaluation. The simulation results show that: After optimization, the path length of the ant colony algorithm is shortened by 16.53% and 16.79% respectively in the scenario with fewer and more obstacles. Moreover, the path quality is effectively improved by reducing the path inflection point, making the path generated by the algorithm more in line with the actual demand. Finally, by setting up a multi-obstacle scenario to simulate the unstructured area of an open-pit mine, the real vehicle simulation test is carried out. The results show that the unmanned mining truck test vehicle equipped with the optimized ant colony algorithm can cross some obstacles, and the passing efficiency in the scene with fewer obstacles is increased by 20.53%, and the passing efficiency in the scene with more obstacles is increased by 31.62%, without any friction with obstacles. Therefore, the proposed parallel 3D path planning method based on optimized ant colony algorithm can effectively shorten the path length, improve the search efficiency and path quality, and give full play to the characteristics of wide body and high underbody of unmanned mining trucks under the premise of ensuring safety. The research results provide a theoretical reference for the development and application of open-pit truck unmanned driving technology.