Under the background of continuously promoting the construction of ecological civilization in Jiangxi Province, analysis of the spatial differentiation of ecosystem service value (ESV) and its influencing factors is of great significance to the protection of ecological environment and the promotion of regional sustainable development. Taking the Nanchang Urban Agglomeration as an example, this paper uses the equivalent factor method and spatial autocorrelation analysis to analyze the spatial distribution characteristics of the average ESV at various grid scales of 1 km×1 km, 3 km×3 km, 5 km×5 km and 10 km×10 km. The influencing factors and scale differences of spatial heterogeneity of ESV at different scales are studied by using geographic detector and spatial regression model. This paper aims to reveal the role of various influencing factors in the Nanchang Urban Agglomeration on the spatial heterogeneity of ESV, and provides scientific basis for improving the quality of ecological environment and promoting green development. The results show that:(1) At different grid scales, the spatial distribution of the average ESV is relatively consistent, which is generally higher in the west and north, and lower in the east and south. There is significantly spatially positive correlation and spatial agglomeration effect at various grid scales, but the agglomeration effect decreases with the increase of grid scale. (2) The results of geographical detectors show that the spatial heterogeneity of ESV in the Nanchang Urban Agglomeration is affected by the synergy of natural environment and economic and social development. The degree of influence of different factors on ESV is significant. At different grid scales, the contribution of HAI is the largest, which is the leading factor affecting the spatial heterogeneity of the Greater Nanchang Metropolitan Area. The reason is that the increase of human activities, such as urban expansion and farmland reclamation, has led to changes in the structure of land use. This affects the spatial distribution of ESV. Moreover, any two factors have higher explanatory power for spatial heterogeneity of ESV than a single factor. However, with the increase of grid scale, the explanatory power of each factor and the coupling and coordination between factors for ESV decreases. (3) With the increase of grid scale, the fitting degree of spatial regression model decreases, and the action intensity and direction of the influencing factors affecting the spatial heterogeneity of ESV are different at different grid scales. At different grid scales, HAI has the strongest explanatory power to ESV, and has obviously negative effects, indicating that human activities have strong inhibition to ESV.