Investigating the spatiotemporal variation of human activity intensity and its determinants is a crucial basis for further revealing the mechanism of human-environment interaction and optimizing the human development mode. In this study, the human activity intensity on the Qinghai–Tibet Plateau (QTP) from 1990 to 2020 was measured based on the quantitative model of land use data and the actual regional background, and the underlying natural and socioeconomic determinants were investigated using spatial econometric methods. The results demonstrate that (1) the human activity intensity in QTP has increased by 11.96%, and there are differences in different spatial scales; the areas with high human activity intensity are distributed in the Hehuang Valley where Xining City and its surrounding areas are located, as well as the One-River and Two-River Area where Lhasa City and surrounding areas are located. (2) Human activity intensity has significant positive spatial spillover, suggesting that local changes will cause changes in the same direction in adjacent areas. (3) The human activity intensity in QTP is affected by various determinants. Concerning socioeconomic factors, the economic level has no significant impact on the human activity intensity in QTP, which differs from the general regional law. Both urbanization and traffic conditions have a significant positive effect, and the impact intensity continues to increase. Concerning natural factors, topographic relief has a significant positive effect; the impacts of temperature and vegetation coverage have changed from insignificant to a significant positive effect; the impacts of precipitation and river network density have not been verified; there is no linear relationship between altitude and human activity intensity in the entire QTP, while it exists in local regions. Finally, this study proposes three policy implications for the realization of a more harmonious human-environment relationship in QTP.