The high accuracy and high resolution of LiDAR data were explored in this study by installing a geomorphometric model of landslides. First, two times of LiDAR data were acquired standing for both before and after a heavy rainfall event. And, the landslides for both times were interpreted. Subsequently, the geomorphometric parameters included slope, aspect, roughness DEM, curvature DEM, OHM, roughness of OHM, and topographic wetness index were acquired through the overlay processes of landslides and the DEM/DSM derived from LiDAR point clouds. The thresholds of every parameter were derived from the statistics of the landslides of whole study area. In addition, some selected non-morphometric parameters were included in the model to cater for all possible features of landslides, such as vegetation index and geological strength. A by-product of volume change estimation also was obtained by the difference of the DEMs before and after the rainfall event. The thresholds of the parameters of landslides are optimized for the model. The overall accuracy predicted by the model is 64.87％. However, the overall accuracy in raster-based evaluation is 64.42％ when buffer zone of old landslides and riverside areas are used. When landslides smaller than 50 m 2 are filtered, the overall accuracy becomes 76.61％ and 72.51％ in 2008 and 2009, respectively. In addition, the change of two LiDAR DEMs obtained before and after Hsiaolin event is reviewed. The largest depth of Hsiaolin Slide is estimated as 84.75 m with a landslide volume of 28.83×10 6 m3. The deepest deposition depth is 85.08 m with accumulated deposits of 17.53×10 6 m3.The deposits in the right-hand bank of Chisan River is estimated as 2.10×10 6 m3. The river bed of Chisan River was raised to a height of 20~22m by the landslide materials.