The efficiency of two sets of Digital Terrain Model (DTM), one based on LiDAR data, and the other on traditional contour lines method were assessed for landslide susceptibility mapping at 1:2000 scale. To evaluate the DTMs, we used the shallow slope stability model, SHALSTAB. The tests were carried out in a basin affected by shallow landsliding caused by extensive rainfall during March 2011, in the urban area of Antonina city (Parana State), southern part of the Serra do Mar mountain range, in Brazil. The geotechnical data needed for the model consisted of two sets of values measured from 2011 landslide scars. In order to validate the landslide susceptibility maps, we compared the spatial pattern of instability classes predicted by SHALSTAB with the mapped landslide scars. The results showed significant difference between the DTMs, especially in the distribution of the most unstable classes.