Evacuation shelters provide shelter and basic living facilities for affected people after disasters. They are an essential part of the urban disaster emergency response system. In this study, a spatialization distribution model of the population in Kumamoto City was constructed using land use data. The P-median model was used to analyze the service areas and the supply and demand of evacuation shelters, the distance of demand points to the corresponding shelters. The results show that the ratio of shelters that met the demand in Kumamoto City is only 41.2%, which is less than half. This ratio increased more in the Urbanization Area. In the Residential Promotion Area and the Urbanization Function Promotion Area with higher population density, the ratio of evacuation shelters with insufficient capacity is higher than that in the Urbanization Area. Among the 15 local hubs in the Urbanization Function Promotion Area, only three shelters have a supply-demand ratio greater than 1. About half of the people can reach the shelter within a distance of 0-1000 meters in terms of evacuation distance. The spatial population estimation model using land use data constructed in this study can simulate the spatial distribution of demand points in evacuation. The supply and demand of shelters were evaluated concerning their capacity. It can be used as a reference for the optimal allocation of evacuation shelters in Kumamoto City and future disaster prevention planning in the Kumamoto compact city.