A novel algorithm based on particle swarm optimization(PSO) of cubic splines is proposed for mobile robot path planning. The path is described by string of cubic splines,thus the path planning is equivalent to parameter optimization of particular cubic splines. PSO is introduced to get the optimal path for its fast convergence and global search character. Experimental results show that a collision-avoidance path can be found fleetly and effectively among obstacles by the proposed algorithm. The planned path is smooth which is useful for robot motion control.