In order to investigate the impact of land use/land cover (LULC) data products of different scales or sources on the estimation of ecosystem service value (ESV), taking the Danjiangkou Reservoir water source area (Hubei section) as an example, two LULC data products were obtained using an upscaling algorithm to generate land use data at 12 different scales ranging from 30 to 3 000 meters. The equivalent factor method was employed to analyze ESV characteristics at different scales, aiming to investigate the spatial scale effect of LULC data on ESV estimation. The results indicate that with increasing spatial scale, the spatial details of LULC data gradually diminish, yet the overall spatial patterns remain relatively stable, and the composition structure of ESV also remains stable. The average difference in ESV estimated by the two LULC data products is approximately 4.23%, with specific differences fluctuating with changes in spatial scale. The ESV estimates at different scales show some differences, but generally remain relatively stable in the range of 30 to 500 meters, while fluctuations are larger at scales of 1 000 meters and above. It is inferred that conducting ESV estimation based on scales of 500 meters and below is more appropriate for the study area. The study results can provide a basis for selecting appropriate scales of LULC data for ESV estimation, and are of reference value for improving the reliability of ESV estimation and promoting the realization of ecological product value.