With the wide area measurement system (WAMS) being widely used in power system, extensive online measured data can be obtained for power system analysis and calculation. In this paper, online parameter identification method of doubly fed induction generator (DFIG) is studied using WAMS data. Firstly, a simplified DFIG model for identification is established, and it is proved that online parameter identification of power system based on WAMS is a closed-loop identification, and this problem can be solved by using prediction error method (PEM). Then, the paper proposes two PEM implementations: Kalman filter method and data transformation method, and compares the similarities and differences between them. Finally, the effectiveness and feasibility of the two methods are validated by using WAMS data obtained from wind farm data in Ningxia Power Grid.