The massive expansion of oil palm plantations has been widely associated with driving deforestation, necessitating accurate spatial inventories to support sustainable land management. This study aims to identify and map the spatial distribution of oil palm plantation land cover in Koto Kampar Hulu District, Kampar Regency, Riau Province, by applying Sentinel-2 satellite imagery and the Random Forest machine learning classification algorithm. The classification model was developed by integrating the original multispectral bands of Sentinel-2 imagery with biophysical index transformations (NDVI, NDWI, and NDBI) to delineate five primary land cover classes: forest, oil palm plantation, settlement, road, and river. Spatial analysis results indicate that the landscape of the study area is predominantly characterized by oil palm plantation cover, encompassing an area of 22,788.56 ha, followed by forest cover at 21,427.95 ha. The Random Forest algorithm proved highly effective in detecting vegetation classes, yielding remarkably low class error rates of 4.96% for oil palm and 3.91% for forest. Variable importance evaluation revealed that Band B2 (Blue) and Band B11 (SWIR) were the most significant spectral predictors in sustaining model accuracy performance. Despite its robustness in classifying vegetation classes, the 10-meter spatial resolution of Sentinel-2 imagery resulted in a high misclassification error rate of 84.92% for small linear objects such as roads, attributable to the mixed pixel phenomenon. Overall, the integration of the Random Forest algorithm and Sentinel-2 optical imagery is strongly recommended for regional-scale oil palm plantation land cover mapping