Aiming at the problem of low accuracy of wind farm equivalent models under fault scenarios, a multi-machine aggregation identification equivalence modeling method for wind farm considering low voltage ride through (LVRT) power characteristics is proposed. Firstly, the wind farm is initially grouped based on the typical differences in active power dynamic characteristics during LVRT of wind turbines. Secondly, the Ng-Jordan-Weiss (NJW) algorithm based on dynamic time warping (DTW) metric is used to realize the secondary division of the cluster, and the final two-stage clustering results are obtained. Then, for the aggregated multi-machine equivalent model of the wind farm, parameter sensitivity analysis is used to determine the key parameters to be optimized. The aggregated values of each parameter are taken as the initial values, and the overall equivalent parameters of the wind farm are optimized using the three strategies of single-machine step-by-step identification, multi-machine sequential identification, and equivalent impedance identification. Finally, the fitting curves and equivalent errors of different methods are compared, and the results show that the proposed method effectively improves the accuracy and adaptability of the equivalent model.