This paper presents a method of applying Particle Swarm Optimization (PSO) algorithm to a flow shop scheduling problem. Permutation encoding of job indices is used to represent particles. One particle of the initial swarm is generated using NEH heuristic [1] and the remaining particles are generated randomly. A continuous swap mechanism is used to improve the performance of the discrete particle swarm optimization (DPSO) algorithm. Performance of the proposed algorithm is evaluated using the benchmark flow shop scheduling problems given by Taillard [2]. The computational results show that the hybrid approach is more effective. © 2006 IEEE.