Parameter identification method is researched for the single-input and single-output systems in this paper. Based on the analysis of method of least squares, the new system identification method based on PID control algorithm is proposed. The PID controller is established by connecting the input and output with the system error and the system parameter. And the results of the simulations state that the new identification method is feasible and effective.System identification is the most important part in modern control theory. The definition of system identification proposed by L. A. Zadeh is that, system identification is determination the system in the gained systems which equivalent to the unknown system, according to the input and output of the unknown system. Traditional identification methods include impulse response, least square and maximum likelihood [1-2]. Identification speed and identification precise are all needed in the system identification. But identification speed is incompatible with identification precision. The improvement of identification do always conduce to the increase of identification precision, it can be verified by the identification examples, which is identified by the traditional identification methods. So for higher identification precision, the identification speed is lower. Now, a lot of high precision identification methods based on neural network or genetic algorithm, is proposed [3-7]. These new identification methods can gain very high identification precision, but it needs too much time. A new system identification method based on PID controller is proposed in this paper. It states by simulation that the new me-thod can improve the identification precision and speed.