Landslides are among the most frequent and destructive geological hazards in Indonesia, particularly in mountainous regions with steep slopes, complex geomorphology, and high rainfall. In addition to their role as natural hazards, landslides act as significant drivers of environmental disturbance and land degradation, highlighting the need for reliable susceptibility assessment as part of environmental monitoring and sustainable land-use planning. This study applies a Random Forest (RF) machine learning model integrated with Geographic Information Systems (GIS) to support environmental monitoring of landslide-prone landscapes in the Tinggimoncong District, Gowa Regency, South Sulawesi, Indonesia. A balanced dataset of 440 sample points (220 landslide and 220 non-landslide locations) was constructed using high-resolution imagery, field verification, and official records, and combined with twelve conditioning factors representing topography, geology, hydrology, land cover, and human activities. Model performance was evaluated using accuracy, F1-score, ROC–AUC, and PR–AUC. The RF model demonstrated high predictive performance (AUC = 0.971; F1-score = 0.95). Feature importance analysis indicates that slope aspect, elevation, slope gradient, and distance to roads are the dominant factors controlling environmentally driven slope instability. The resulting probabilistic and classified susceptibility maps identify zones of high environmental instability that spatially correspond with areas prone to land degradation due to recurrent landslides. The proposed RF-based framework provides a robust basis for environmental monitoring, land degradation management, and sustainable spatial planning in complex tropical terrains.