This paper focuses on the development of intelligent path planning algorithm for unmanned vehicles automatic guidance such as the shortest path, the stable path in a terrain with a single destination and multiple waypoints. The experiment results show that the proposed algorithm can not only reduce the uncertainty of the evolutionary computation caused by the probability model, but also avoid the problem of particles falling into the local optimal area effectively. The proposed algorithm for multiple waypoints path planning is also more efficient than other route planning methods.