This paper analyzes the distribution of landslides in southern Taiwan due to typhoon or rainstorm from 2009 to 2011. The genetic adaptive neural network (GANN) integrated with texture analysis and geographic information system (GIS) is applied to capture disaster records and surface information for interpretation and classification of satellite images, and then analysis of land use change. The multivariate hazards evaluation method (MHEH) is employed to quantify the weight of the disaster factors, and then build the landslide potential model. Furthermore, this paper also explores the correlation of rainfall events on slope land development, scale and regional characteristics. The results of this study will provide a reference for the development of subsequent slope development strategies and the evaluation of landslide.