The main objective of this study is to compare the results of decision tree classifier and its ensembles for landslide susceptibility assessment along the National Road 32 of Vietnam. First, a landslide inventory map with 262 landslide locations was constructed using data from various sources that accounts for landslides that occurred during the last 20 years. Second, ten landslide conditioning factors (slope, aspect, relief amplitude, topographic wetness index, toposhape, distance to roads, distance to rivers, distance to faults, lithology, and rainfall) were prepared. Third, using decision tree and two ensemble techniques i.e. Bagging and AdaBoost, landslide susceptibility maps were constructed. Finally, the resultant landslide susceptibility maps were validated and compared using a validation dataset not used during the model building. The results show that the decision tree with Bagging ensemble technique have the highest prediction capability (90.6 %), followed by the decision tree (87.8 %) and the decision tree with AdaBoost (86.2 %).