Abstract: The remote mountainous areas at the end of the distribution network are often regarded as weak areas of the distribution network due to insufficient power supply reliability. In order to improve the power supply reliability of such areas, the concepts and methods of virtual substation and cluster planning are introduced, and a bi-level location and capacity planning model based on virtual substation for distributed power sources and energy storage is proposed. The upper-level model takes the largest cluster efficiency index as the goal, and divides the weak areas of the distribution network into clusters from both structural and functional aspects, whose planning result is the division plan of each cluster. The lower-level model aims to minimize the annual comprehensive cost considering the user’s loss costs, and build virtual substations within each cluster, whose results are the access location and capacity of distributed power sources in the cluster, and the access capacity and power of energy storage devices. Focusing on the characteristics of the cluster planning model, a cost calculation method for users’ power outage loss considering island operation is proposed, and an iterative hybrid particle swarm algorithm is used to solve the problem. A typical scene with distributed small hydropower stations in Yunnan is taken as an example to verify the feasibility of the proposed model and the effectiveness of the solution method.