With the highly penetration of large-scale wind farms (WFs), accurate grid-connected WF models are crucial to be developed for studying the stability of power grids. Due to the large scale of the detailed model, a grid-connected WF is normally simplified into several aggregated wind turbines (WTs) with the dynamic equivalent technique. However, the identification of large amounts of parameters in an aggregated WT is one main issue that reduces the accuracy of the developed models. Moreover, the number of the aggregated WT models which need to be identified simultaneously further complicate the issue. Therefore, in the paper, a dynamic equivalent modeling approach which could estimate the full parameter of the equivalent WF model is especially developed. First, a WT clustering approach is designed to simplify the WF model into several aggregated WTs. Then, the external system of the aggregated WTs are simplified with the hybrid dynamic simulation (HDS) technique, by which the parameters of each equivalent WT could be identified independently. Also, the WT parameters are grouped with the trajectory sensitivity and identified using the multistage parameter identification approach. The simulation results demonstrated the effectiveness of the proposed approach.