In recent years, grassland in Qinghai province has been affected by climate change and variously impacted by a range of human activities, sometimes causing degradation and sometimes facilitating restoration. Here we report a survey of the spatial and temporal variation in grassland productivity across Qinghai Province for the period from 2001-2017. To extract data on grassland net primary productivity (NPP, MOD17A3HGF.006), Sen’s slope together with the Mann-Kendall test (Sen-MK), correlation analysis and residual analysis were used as decision nodes of a decision tree. Then we used the tree to evaluate grassland NPP time trends and their spatial heterogeneity during the study period and identify associated factors that may be change drivers. The results were as follows: 1) There were 11.41×104 km2 of grassland with significant (P<0.05) NPP trends in Qinghai Province during the 2001-2017 period. The areas of extremely significant (P<0.01) increase, significant (0.01≤P<0.05) increase, significant decrease and extremely significant decrease were respectively 11.88%, 17.25%, 0.20% and 0.08% of the grassland in the province. Remarkably, the overall change is a rise in NPP. 2) For the regions where grassland NPP trends were significant, the proportion explained by temperature was the largest (60.66%), followed by human activities (23.45%) and human activities + temperature (9.49%). Hence, temperature and human activities were the major factors driving significant grassland NPP trends. 3) Temperature, precipitation and human activities mainly had positive impacts on NPP. Human activities were the dominant factor linked with grassland degradation and were associated with 77.49% of the grassland area with a decreasing NPP trend. However, human activities have also achieved remarkable results in grassland restoration and ecological protection projects in Qinghai Province.