Abstract: In response to the issues of inadequate utilization of geographic information and insufficiently balanced zoning load in traditional substation planning, a method for substation siting and capacity determination integrating geographic information system (GIS) and Voronoi-Kmeans hybrid zoning is proposed. Firstly, the land price weight matrix is generated based on the spatial power load and actual land information based on GIS. Secondly, based on the results of spatial power load gridding, the hybrid Voronoi-Kmeans algorithm is designed to partition the spatial power load. Subsequently, the substation capacity is determined based on the total load of each zone and the capacity ratio. Finally, an annual cost-minimization planning model for substations that incorporates land prices is constructed and solved using an improved genetic algorithm. The analysis of the arithmetic example demonstrates that the proposed site selection and capacity determination method improves the accuracy of substation planning and reduces the planning cost of the power system.