In order to improve the orderly control and flexible adjustment capabilities of large-scale distributed power sources connected to the distribution network,it is currently an effective method to reasonably divide the distributed power sources into clusters and use the cluster as the basic unit for regulation. To this end,a distributed power cluster partitioning method that considers the adjustable capacity and response speed of flexible resources is proposed. Firstly,the modularity index and the active power balance index are proposed to characterize the structural and power balance capability of clusters. Secondly,based on the analysis of the cluster's flexible response capability,and in view of the problem that it is difficult to quantify the impact of the flexible resource response capacity and response speed on the adjustment capability,it is proposed to take the cumulative area of deviation between the response curve of cluster output and the net load curve as the indicator of the flexible response capability of clusters,and the influence of the allocation of energy storage capacity on the delineation of clusters is also taken into account. Then,a cluster division objective function that comprehensively considers the modularity index,active power balance index,cluster flexible response capability index and energy storage unit allocation index is proposed,and this function is solved using an improved adaptive genetic algorithm. Finally,the proposed method is verified in the improved IEEE 33-node system and IEEE 69-node system. The results show that the cluster partitioning method based on the proposed comprehensive performance index can improve the cluster flexible response capability by more than 20% on the basis of ensuring the structural strength of the cluster,which verifies the effectiveness and superiority of the cluster partitioning method proposed in this paper.