Best management practices (BMPs) are management operations that reduce pollution and improve water quality. This study assessed the spatial variability of BMPs effectiveness within the Yazoo River Watershed (YRW) using Soil and Water Assessment Tool (SWAT). Two field-scale watersheds, Merigold Watershed (MW) from the Delta and Skuna River Watershed (SRW) in the Bluff Hills, were selected within the YRW. The SWAT model was calibrated and validated for monthly streamflow, and daily total suspended solids (TSS), total nitrogen (TN), and total phosphorus (TP). Monthly evaluated streamflow Nash–Sutcliffe efficiency indices were in the range from 0.60 to 0.86; daily evaluated TSS indices were in the range from 0.11 to 0.15, TN from 0.11 to 0.12, and TP from 0.05 to 0.26 during model calibration and validation periods. BMPs were implemented in MW and SRW to analyze the spatial variability effect on water quality. Cover crops (CC), vegetative filter strips (VFS), and a combination of VFS and CC were applied as BMP scenarios. Overall, a larger reduction in streamflow was about 15%, sediment about 26%, and nutrient loads, which was about 39% (TN) and 50% (TP), was determined in the MW, whereas in the large watershed (YRW) the reductions in streamflow, sediment, TN, and TP loads were about 4%, 5%, 30%, and 24%, respectively. Therefore, the novelty of this research is to compare the efficiency of a BMP in pollutant reduction at two different geographic scales. The results from this study could help farmers, scientific researchers, precision management planning, and implementing agencies select appropriate BMP for field-scale water resources management in minimizing sediment and nutrient quantities in surface water.