Alpine grasslands on the Qinghai-Tibetan Plateau are sensitive and vulnerable to climate change and human activities. Climate warming and overgrazing have already caused degradation in a large fraction of alpine grasslands on this plateau. However, it remains unclear how human activities (mainly livestock grazing) regulates vegetation dynamics under climate change. Here, alpine grassland productivity (substituted with the normalized difference vegetation index, NDVI) is hypothesized to vary in a nonlinear trajectory to follow climate fluctuations and human disturbances. With generalized additive mixed modelling (GAMM) and residual-trend (RESTREND) analysis together, both magnitude and direction of climatic (in terms of temperature, precipitation, and radiation) and anthropogenic impacts on NDVI variation were examined across alpine meadows, steppes, and desert-steppes on the Qinghai-Tibetan Plateau. The results revealed that accelerating warming and greening, respectively, took place in 76.2% and 78.8% of alpine grasslands on the Qinghai-Tibetan Plateau. The relative importance of temperature, precipitation, and radiation impacts was comparable, between 20.4% and 24.8%, and combined to explain 66.2% of NDVI variance at the pixel scale. The human influence was strengthening and weakening, respectively, in 15.5% and 14.3% of grassland pixels, being slightly larger than any sole climatic variable across the entire plateau. Anthropogenic and climatic factors can be in opposite ways to affect alpine grasslands, even within the same grassland type, likely regulated by plant community assembly and species functional traits. Therefore, the underlying mechanisms of how plant functional diversity regulates nonlinear ecosystem response to climatic and anthropogenic stresses should be carefully explored in the future.