<p>Changes in land use and land cover in ecologically sensitive regions such as Karnataka’s Western Ghats have significant implications for environmental sustainability, urban expansion, and natural hazard risk. This study aims to assess LULC transformations from 2010 to 2023 using the GeoML-LULC Classification and Assessment Framework (GeoML-LCAF), a machine learning-based geospatial framework developed on the Google Earth Engine system. The framework combines multitemporal satellite data from Sentinel-2, MODIS, and Dynamic World with NDVI, slope, aspect, curvature, and lithology geospatial parameters extracted from DEMs. These layers were used as input to train three different machine learning classifications Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosting (GF) for LULC classification. The data was layered and then processed in GEE for feature extraction. The findings indicate a notable decrease of 35 percent in forest coverage, alongside a 15.73 percent rise in urban zones and a 7.22 percent growth in farm land. The Gradient Boosting model outperformed the rest of the evaluated models showing a classification accuracy of 94.34% and the Kappa coefficient of 0.93. This superb agreement with ground truth data means that your model is performing very well. These findings illustrate the increasing anthropogenic stresses placed on the region’s landscape and illustrate the value of applying machine-learning based geospatial approaches for informed land use decision-making and sustainable land development in landslide risk zones.</p>