Understanding environmental changes and promoting the sustainable management of watersheds requires long-term monitoring of land use and land cover (LULC), especially in Mediterranean regions subject to high climate variability and growing anthropogenic pressure. Using the whole Landsat imagery archive accessible via the Google Earth Engine platform, this study introduces a novel spatiotemporal land use and land cover dataset (LULC_AllalElFassi) for the Allal El Fassi watershed spanning the years 1985–2025. The methodology ensures the geographical and temporal coherence of the produced maps while detecting changes in land use and land cover by combining a locally adaptive classification strategy with the Continuous Change Detection and Classification (CCDC) algorithm. Eight land use and land cover classifications could be mapped at a spatial resolution of 30 m thanks to the dataset, which was created using 2,000 training samples and 800 independent validation points. The total accuracy of the results was 82.82%. Water bodies (93.65%), built-up areas (92.65%), agricultural land (87.64%), and dense woods (87.81%) had especially good categorization accuracy. Long-term landscape dynamics investigation showed a significant increase in open forests (+898.50%), a notable expansion of agricultural land (+10.38%) and built-up areas (+789.63%), and a decrease in thick forests (−9.35%) and herbaceous and poorly vegetated regions (−4.73%). These alterations show that the forest canopy is gradually opening, agricultural production is becoming more intensive, and urbanization is increasing across the watershed. The suggested dataset is a useful decision-support tool for hydrological modeling, water erosion assessment, forest resource management, and sustainable planning of Mediterranean watersheds. It also provides a reliable data source for long-term environmental monitoring.