Abstract: The identification on parameters of closed loop time delay systems has been an important topic in advanced industrial control field. However, due to the existence of time delays, the controlled variable cannot timely reflect the external disturbance, which may result in obvious overshoot and poor stability. Using the advantages of particle swarm optimization algorithm and chaotic motion, this paper proposes a direct identification method on the transfer function of closed loop time delay systems, rather than transforming the transfer function to the state equation. Using z transform, the transfer function of closed loop time delay system is changed into discrete differential equation. Besides, the time delay is approximated by means of first order Pade approximation. CPSO method is adopted to minimize the error criterion function so as to obtain the estimation of the model parameters. It is shown from the simulation results that the proposed method has fast convergence speed and higher identification accuracy.