Particle swarm optimization (PSO) has shown its good search ability in many optimization problems. However, PSO often easily fall into local minima because the particles could quickly converge to a position by the attraction of the best particles. Under this circumstance, all the particles could hardly be improved. This paper presents a hybrid PSO, namely LSPSO, to solve this problem by employing an adaptive local search operator. Experimental results on 8 well-known benchmark problems show that LSPSO achieves better results than the standard PSO, PSO with Gaussian mutation and PSO with Cauchy mutation on majority of test problems.