Landslides are the most common and devastating natural calamities in the Darjiling Himalaya. It cripples the daily life of inhabitants for three to four monsoonal months of the year. Hence, the present study aimed to compare the predictive performance of two bivariate models, namely, the frequency Ratio (FL) and fuzzy Logic (FL), for assessing the susceptibility of landslides in the region. A total of 1556 landslide locations were identified in the study area from the inventory reports and then verified with Google Earth images and field surveys. Of these 1556 identified landslide points, (70%) 1089 were randomly used for data training, and the remaining (30%) 467 points were used for data validation. Fifteen landslide-causing parameters, namely, elevation, slope, rainfall, distance to river, distance to lineament, distance to road, aspect, geomorphology, curvature, plan curvature, profile curvature, seismic zone, LULC, TWI (topographic wetness index) and NDVI (Normalized Difference Water Index), were used as independent variables in the modeling process and to construct the spatial database. Subsequently, geographical information systems (GIS) and two predictive models were ensembled to assess landslide susceptibility. The resulting landslide susceptibility maps (LSMs) were subsequently classified into four classes, namely, low, medium, high, and very high susceptibility to landslide occurrence, using the natural breaks classification method in the GIS environment. Finally, validation results obtained from the receiver operating characteristics (ROC) curve showed that the landslide susceptibility map constructed with the FR model (76.1%) was slightly more accurate than that produced with the FL model (73.6%). Landslide susceptibility assessments using comparative predictive models are always better than those using any single model. The findings of this study could lead to potential resources for future planning and development in the study area.