Particle swarm optimization (PSO) method was proposed by Kennedy and Eberhart in 1995, which can be used to solve a wide array of different optimization problem. The PSO idea is inspired by natural concepts such as fish schooling, bird floc king and human social relations. Some experimental results show that PSO has gre ater “global search” ability, but the “local search” ability around the opti mum is not very good. In order to enhance the “local search” ability of the traditional PSO, two improvement methods for the PSO, that is, PSO with simulated annealing (PSOwSA) and PSO with division of work (PSOwDOW), are introduced by analyzing deeply the traditional PSO. Experiments for several benchmark problems s how that PSOwSA and PSOwDOW can overcome the fault of traditional PSO and increa se the optimization power of the particle swarm.