[Objective] Clarifying the changes in net primary productivity（NPP）and the driving factors is of great significance to measure the carbon sequestration capacity of vegetation and support carbon sink trading. [Methods] Net primary productivity data were obtained based on MODIS-NPP data，and combining the Theil-Sen Median trend analysis，Mann-Kendall test，and the characteristics of spatiotemporal variation of NPP in Kunming from 2001 to 2020 were revealed. The optimal parameters-based geographical detectors（OPGD）model was adopted to explore the fluencing mechanisms of climate，topography and anthropogenic factors on NPP. [Results] 1）The annual mean distribution of vegetation NPP in Kunming City was spatially and temporally heterogeneous，with a distribution pattern higher in the west and lower in the east，and increased significantly at an average annual growth rate of 7.08 g/（m²·a）（C-based）. 2） The fragmentation degree of the vegetation landscape increased；the fragmentation degree was mismatched with vegetation NPP in terms of space，and it had a significant negative effect on the spatial and temporal distribution of vegetation NPP. 3）Changes in vegetation NPP were the result of the combined effects of climatic， anthropogenic and topographic factors； the kernel Normalized Difference Vegetation Index（kNDVI），land surface temperature，and vegetation landscape fragmentation had stronger explanatory power than factors such as temperature，precipitation，and slope；and the interactions between the factors better explained the spatial and temporal changes in vegetation NPP. [Conclusion] The vegetation NPP in Kunming showed an increasing trend from 2001 to 2020 in general. However，the ecological situation in certain areas of the main urban districts is not optimistic. In the future，comprehensive planning should be implemented to scientifically promote the coordinated development of ecological protection and social economy.