Landslides are common natural hazards in Malaysia. These landslides can be systematically assessed and mapped through traditional mapping frameworks using geoinformation technologies (GIT). The aim of this study was to apply, verify, and compare an artificial neural network model and its cross application of weights for landslide susceptibility analysis in three Malaysian study areas, namely, Penang Island, Cameron Highland, and Selangor, using a geographical information system (GIS). Landslide locations were identified in the study areas from interpretation of aerial photographs, field surveys, and inventory reports. The landslide-related spatial database was constructed from topographic, soil, geologic, and land-cover maps. The 11 factors that influence landslide occurrence were extracted from the database, and the weight of each factor was computed. Different training sites were selected randomly to train the neural network, and nine sets of landslide susceptibility maps were prepared. Landslide susceptibility maps were drawn for the study areas using weight derived not only from the data for that area, but also using that of each of the other two areas (nine maps in all) as a cross-check of method validity. The verification results show that among the nine cases, the best accuracy (83.99 percent) was obtained in the case of the Cameron-based Cameron weight, whereas the Penang-based Cameron weight showed the worst accuracy (70.58 percent).