Abstract: Such disadvantages as sensitivity to inertia coefficient, time consuming etc. are always encountered in the application of particle swarm optimization algorithm. The procedure used in harmony search algorithm and the assumption of discontinuous flying are imported into the particle swarm optimization algorithm. Firstly, when the boundaries are exceeded, the harmony search algorithm is implemented to obtain a new trial solution; and secondly, random number of particles instead of static number are allowed to update their velocities and positions in order to decrease the time consumed by the algorithm. The particle swarm optimization algorithm and the mixed version of particle swarm optimization algorithm are applied to the determination of critical slip surface of given complicated soil slopes. The comparative study shows that the mixed version of particle swarm optimization algorithm tends to find better solution spending little time than particle swarm optimization algorithm.