Landslide susceptibility mapping (LSM) is essential for reducing landslide risk in tectonically unstable environments such as Idukki in the Western Ghats, India. To address the uncertainty associated with reliance on a single predictive approach, this study integrates knowledge-driven Weighted Multi-Criteria Decision Analysis (WMCDA) with six data-driven machine learning (ML) classifiers. A geospatial database was constructed using six primary conditioning factors, including a Digital Elevation Model (DEM), mean monsoon rainfall (2015–2020), Land Use/Land Cover (LULC), lithology, and proximity to roads and waterways. Terrain attributes, such as slope, aspect, Topographic Wetness Index (TWI), and curvature, were derived from the DEM and incorporated into the ML models. Model performance was evaluated using a 70:30 train-test split with 5-fold cross-validation based on a landslide inventory of 40 events and an equal number of non-landslide samples. Among the evaluated classifiers, the multilayer perceptron (MLP) achieved the best overall performance, with a Receiver Operating Characteristic – Area Under the Curve (ROC-AUC) of 0.904, a Precision-Recall – Area Under the Curve (PR-AUC) of 0.917, and a Brier score of 0.137. Feature importance analysis identified proximity to roads as the most influential factor, with a score of 0.178, highlighting the substantial role of anthropogenic disturbance in slope instability. Density-based validation further showed that high and very high susceptibility zones, covering 24.3% of the study area, captured 82.5% of historical landslide events. These findings provide a robust basis for hazard zonation, land-use planning, and infrastructure decision-making in monsoon-affected mountainous regions.