In this paper, a new dynamic clustering method for the DDPMSG-based wind farm is proposed based on the action time of current limiting component in the grid-side controller. Firstly, the affecting factors on the action time of current limiting component in grid-side controller are analyzed, and it is pointed out that the action time of limiting component dominates the dynamics of DDPMSG-based wind turbines. Secondly, the estimating method of limiting component action time is proposed based on the steady active power and the terminal voltage sag. Finally, the DDPMSG-based wind turbines are grouped based on the limiting component action time with K-means method. The 10-machine 39-node system integrated with a wind farm is used as the test system, and the dynamic responses of the wind farm based on its detailed model and the equivalent one are compared. Meanwhile, the small-signal stability and transient stability of the overall system are analyzed with detailed wind farm model and equivalent wind farm model. Compared result validates the clustering method proposed in this study.