Alpine grassland is sensitive to climate change, and many studies have explored the trends in alpine vegetation. Most research focuses on the effects of climate warming and increased humidity on vegetation greening. However, less attention has been given to the positive impacts of human activities, particularly ecological restoration projects (ERPs). Our study utilized the CASA (Carnegie Ames Stanford Approach) model to simulate the net primary productivity (NPP) of alpine grasslands on the Tibetan Plateau (TP) from 2000 to 2020. Additionally, a moving window approach was employed to comparatively analyze the changes in the response characteristics of NPP to climate change before and after the implementation of ERPs. Our results indicated: (1) The NPP exhibited a fluctuating upward trend. The NPP growth rates of alpine meadow, alpine grassland, and desert grassland were found to be 2.38, 1.5, and 0.8 g C·m⁻²·a⁻¹, respectively. (2) The annual average NPP and annual growth rate of alpine grasslands after the implementation of ERPs were both higher than before, indicating that ERPs have intensified the growth trend of NPP in alpine grasslands. (3) ERPs have reduced the responsiveness of alpine grassland NPP to temperature variations and enhanced its responsiveness to changes in precipitation. In detail, ERPs enhanced the responsiveness of NPP in alpine meadow to both temperature and precipitation, reduced the responsiveness of NPP in alpine steppe to temperature while enhancing its responsiveness to precipitation, and mitigated the changes in the response of NPP in desert steppe to temperature and significantly enhanced its responsiveness to precipitation.