The basic particle swarm optimization algorithm(PSO) is easy to fall into local minima and convergence slowly, so an improved PSO algorithm with two mutations based on normally throwing distribution points in the updating process is presented. One of the mutations is used to enhance the local searching ability, the other is used to increase the ability of finding the global optimum and avoid all particles falling into a neighborhood of a local minima. Experimental results show that the global search capability of the proposed algorithm is improved significantly and the convergence speed is faster.