Abstract: For the problem of leakage identification of water distribution networks, this paper proposed a leakage identification method based on virtual pressure partition. The dimension of the leakage identification problem was reduced by virtual pressure partition of water distribution networks, and the distribution of real leakage flow in the network was optimized by a genetic algorithm. The method was divided into virtual pressure partition and leakage zone identification. Firstly, based on the improved K-means clustering algorithm, the network nodes were divided into different regions with adjacent locations, similar average free water heads and connected topological structures. Secondly, the regional proportional coefficient was taken as the decision variable, and the objective function was to minimize the difference between the simulated value and the measured value at the pressure measuring point, and the real leakage situation of each region was identified. A small verification case was designed based on the No. 8 area of the water distribution network in L Town, Z City, to verify the accuracy of the method in identifying different leakage conditions. Subsequently, the water distribution network in L Town was also used for practical engineering application. The results show that the proposed method divides the nodes well, and the identification results of the serious leakage area in the network are consistent with the actual situation, which can assist in the leakage detection of large water distribution networks.