Understanding land-use change (LUC) and its influence on landscape structure is necessary for sustainable watershed management. This study integrates a novel multi-temporal remote sensing and landscape metrics (LM) based approach to investigate long-term LUC dynamics and their geomorphic implications in the Cham Gardalan watershed, western Iran, from 1990 to 2023. Fragstat 4.2 was employed to derive structural and functional LM across five land-use classes (forest, rangeland, orchard, barren land, water bodies). Landsat imagery was processed through Google Earth Engine to generate consistent land-use classifications. Significant forest loss (from 118.14 km2 to 79.41 km2) and expansion of barren land (388.54 km2) indicate intensified degradation and increased soil erosion risks. Temporal trends in patch density, edge density, and contagion metrics indicate increasing fragmentation and reduced ecological connectivity through 2010, followed by partial recovery linked to conservation and reforestation efforts. Correlation analysis among LM further highlights that forests exhibit the highest spatial contrast with barren areas, orchards, and water (negative correlations ranging from-0.76 to-0.98) and a positive correlation with pastures (0.68-0.89), verifying the detrimental impacts of human activities on sediment control and landscape cohesiveness in the region. Integrating LM with multi-temporal satellite data provides a replicable, quantitative framework for assessing the spatial evolution of watershed degradation. By analyzing structural changes induced by LUC, this study aims to better understand landscape dynamics in semi-arid environments. The framework introduced can be adapted to other geomorphologically vulnerable regions to support soil conservation and sustainable land-use planning.