Particle swarm optimization algorithm (PSO) is used to solve optimization problems. In this algorithm, communication between particles is an important factor that enables us to find an optimum or near optimum solution. In parallel version of this algorithm, communication is ensured by the migration of particles between sub-population. In this study, we present a new migration pool technique when selecting the particle to migrate. According to the number of processors, we set three migration pool size. Then, we apply the proposed migration pool technique on 9 different test functions. The results show what the migration pool size should be for each function to reach the optimal solution.