In this study, the effect of NDVI is observed on the landslide susceptibility maps (LSMs) being created using GIS-based bivariate and multivariate statistical models such as frequency ratio (FR), information value (IoV), multiple linear regression (MLR), and logistic regression (LR) based on the 108 training and 56 testing landslide points and ten causative factors. The accuracy of the LSM produced was highest for the IoV model with and without considering the NDVI as a causative factor; however, the accuracy of each of the models was increased when NDVI was considered, i.e., for the FR model, the AUC was increased from 0.85 to 0.92, while for MLR model the AUC value increased from 0.86 to 0.90 and0.84 to 0.92 for LR model. Additionally, the distances closer to the road and streams, jointed and fragmented rock strata, western direction aspect were more likely to experience landslides. We, therefore, suggest using NDVI as a causative factor in the LSM preparation while working in dense vegetation and tropical areas.