This paper studies an accurate identification method for the parameters for a class of linear system, firstly uses PSO (Particle Swarm Optimization) to optimize and iterate the model, chooses suitable deviation standard as the fitness function of the PSO algorithm to iterate the relative parameter speed and size of each particle, on the basis of this, seeks the particle with minimum fitness vale, deducts the optimal fitness function value, implements realtime and accurate estimation of systematic parameters and finally uses experiment to test the accuracy and effectiveness of PSO algorithm based parameter identification method.