The sensitivity of vegetation productivity to precipitation (S_ppt) is a key metric for understanding the variations in vegetation productivity under changing precipitation and predicting future changes in ecosystem functions. However, a comprehensive assessment of S_ppt over all the global land is lacking. Here, we investigated spatial patterns and temporal changes of S_ppt across the global land from 2001 to 2018 with multiple streams of satellite observations. We found consistent spatial patterns of S_ppt with different satellite products: Sppt was highest in dry regions while low in humid regions. Grassland and shrubland showed the highest S_ppt, and evergreen needle-leaf forest and wetland showed the lowest. Temporally, S_ppt showed a generally declining trend over the past two decades (p < 0.05), yet with clear spatial heterogeneities. The decline in S_ppt was especially noticeable in North America and Europe, likely due to the increase in precipitation. In central Russia and Australia, however, S_ppt showed an increasing trend. Biome-wise, most ecosystem types exhibited significant decreases in S_ppt, while grassland, evergreen broadleaf forest, and mixed forest showed slight increases or non-significant changes in S_ppt. Our finding of the overall decline in S_ppt implies a potential stabilization mechanism for ecosystem productivity under climate change. However, the revealed S_ppt increase for some regions and ecosystem types, in particular global grasslands, suggests that grasslands might be increasingly vulnerable to climatic variability with continuing global climate change.