The evolution of landscape patterns due to land use change and urbanization has a great impact on ecosystem services, especially on food production services. As natural ecosystems, exploring the spatial and temporal influence mechanisms of watershed landscape patterns on food production services can enhance the supply of food production services and maintain food security through landscape pattern optimization. This study took Jiangxi Province as the study area, and measured the spatial distribution of grain production services for five periods in 2000, 2005, 2010, 2015, and 2020; from three dimensions of shape, area, and spatial distribution, we selected the patch density (PD), the mean patch size (MPS), the patch cohesion index (COHESION), and the Landscape Shape Index (LSI) to measure the landscape pattern characteristics of the watershed unit; using spatio-temporal geographically weighted regression model to explore the influence mechanism of landscape pattern changes on grain production services; and based on the spatial heterogeneity of the influence mechanism to divide the landscape regulation zones, and put forward the proposal of zonal regulation of the landscape pattern. The results show that: 1) From 2000 to 2015, the overall supply of food production services in Jiangxi Province showed a fluctuating upward trend, and then fell back in 2015, with large differences in the capacity of food production services in various basins, and the center of gravity extended from the Poyang Lake Plain to the southwest, forming the aggregated distribution characteristics centered on the plains and basins, and the GP presented the pattern of Poyang Lake Plain > Jitai Basin > Gannan Hills > Northwestern Low Hilly Area. 2) During the period of 2000-2020, the landscape pattern was changed to a more heterogeneous landscape, and a proposal for zoning control of the landscape was made based on the influence mechanism.) During the period of 2000-2020, the direction and degree of change of landscape pattern index differ greatly, with the patch density and landscape shape index increasing by 5.35% and 1.39% respectively; the average patch area and landscape connectivity index decreasing by 5.14% and 0.68% respectively, and the landscape pattern tends to be complicated, fragmented and dispersed, and the landscape fragmentation is roughly presented as The landscape fragmentation is roughly in the distribution pattern of “high in the plains and basins, low in the mountains and hills; high in the north of Gan, low in the south of Gan”. 3) The spatial and temporal evolution of GP is the result of the joint action of many factors, and the influence of each landscape driving factor on GP shows a piecewise or bandwise spatial distribution pattern. Among them, cultivated land area, water density, and water shape have positive effects on GP, while water area, woodland density, and grassland shape have negative effects on GP.4) Summarizing the spatial differentiation characteristics of landscape factor coefficients, the study area is divided into four clustering areas, namely, northeast Gan monsoon area, Poyang Lake plain area, central main production area, and southern hilly area, and corresponding regional landscape regulation suggestions are proposed. The results of the study can provide reference for the enhancement of food production services and landscape pattern regulation in Jiangxi watersheds, and strengthen the pertinence of the regulation measures in the research unit.