Net primary productivity (NPP) is an important parameter for evaluating ecosystem vegetation growth and a critical indicator for representing the quality and function of terrestrial ecosystems. This paper collected the MODIS NPP data, DEM, meteorological, hydrological, and human activity data to reveal the spatiotemporal dynamic characteristics and driving factors of NPP in the Qilian Mountains (QLM) during the last 20 years. Based on the pixel and county scales, we investigated the spatiotemporal variation of the NPP in the QLM by employing spatial analysis and trend analysis. Furthermore, we studied the response of NPP to the meteorological factors (i.e. annual average temperature and annual precipitation) by using partial correlation and regression analysis. The Geodetector model was employed to reveal the driving mechanism of the NPP change. At last, the Hurst index was introduced to predict the future change trend of NPP. The results indicated that there was a fluctuant increase trend of NPP in the QLM during 2000-2020 with a slope of 2.38 g C m^-2 a^-1, the rising degree of the cultivated vegetation and broad-leaved forest was larger compared with other vegetation types. In the past 20 years, 75.37% of the regional NPP increased in the pixel scale, which primarily located in the southeastern regions of the QLM. At the county level, the NPP growth rates of Gulang County, Ping'an County, Hualong County, and Yongdeng County were fast, while that of Qilian County, Haixi County, Delingha County, and Menyuan County were slow. The spatial distribution of NPP in the QLM has obviously spatial agglomeration characteristics. The high-value agglomeration area mainly located in the southeast regions of the QLM, while the low-value agglomeration area mainly situated in the northwest regions of the QLM. The increase in annual average temperature and precipitation promoted the increase of NPP, but the responses of NPP to temperature and precipitation were significantly diverse in different regions. Precipitation, saturated water vapor pressure deficit, and evapotranspiration were the main driving factors of NPP change, and there were interactions effects of driving factors on NPP change, which were two-factor enhancement and nonlinear enhancement effects, respectively. The changing trend of the NPP in the QLM will dominate by the increase with nonpersistent in the future, which indicates that vegetation development in the QLM will face greater uncertainty. The results can help to reveal the nonlinear response mechanism of vegetation NPP to climate change and human activities under the background of global climate change and provide scientific support for the ecological protection and sustainable development of the QLM.