Abstract: In order to effectively improve the fitting precision of the equivalence model to the wind farm external characteristics,a dynamic equivalence method of wind farms with permanent magnet synchronous generator（PMSG）based on improved D-K clustering algorithm is proposed. Three critical steps are involved in this method. First,researching the clustering principle,the low voltage ride through（LVRT）capability of a PMSG is taken into account. The chopper circuit conduction is selected as the clustering basis,and the terminal voltage of each PMSG during the fault is used as the clustering index. Second,clustering the wind farm,an improved D-K clustering algorithm is first applied to cluster the wind farm. The innovative algorithm can improve the accuracy and stability of the traditional K-means algorithm,solve the problem of determining the most appropriate number k of clusters and best initial centers.Third,calculating equivalent parameters,PMSGs of the same cluster are equivalent as one PMSG to realize the dynamic equivalent modeling of wind farm. The simulation results show that the proposed equivalent model can reflect the dynamic and transient operating characteristics of the wind farm more accurately compared with the traditional equivalent model. Thus,the model can meet the requirements of the transient analysis in power systems.