In this paper, the performance of the particle swarm optimization(PSO) algorithm is studied from the system dynamics point of view. The dynamics of the particles in PSO algorithm are considered as second-order systems. Depending on the selections of the parameters, the second-order systems have over-damped, critically damped, underdamped, or undamped responses. Different responses give the algorithm different types of performance. Therefore, in this paper, we derive the conditions for parameters in the PSO algorithm such that the particles have different responses. The exploration and exploitation of PSO are discussed numerically. Moreover, due to the fact that the discrete model of PSO is converted from a continuous model by certain sampling ratio, the sampling ratio variable is introduced to the PSO algorithm. With different sampling ratios, the stability region of the PSO algorithm is increased and the performance of the algorithm is changed. Numerical examples are provided to demonstrate the performance of the PSO algorithm with different selections of the parameters.