Even though detailed building inventory data are necessary for estimating earthquake damage reliably, most developing countries do not have sufficient data for such estimations. This necessitates a way for finding building distribution and feature easily. In this study for estimating the number of households in all building categories of different structures or floor numbers in Lima, Peru, where a great earthquake is expected, we propose an estimation method based on existing GIS data from a census, satellite imagery, and building data from field surveys, and apply it to estimate the entire area of Lima for create building inventory data. Building fragility functions were used to calculate a severe damage ratio of buildings due to the expected earthquake. The rate was multiplied by created building inventory data to estimate the number of households in damaged buildings. Furthermore we clarified damage reduction by retrofitting for low earthquake-resistant buildings.