Abstract: Improving green total factor productivity (GTFP) is an important initiative to promote high-quality economic and social development in the Qinghai-Tibet Plateau region. Based on the panel data of cities in the Qinghai-Tibet Plateau from 2000 to 2019, the spatial-temporal evolution characteristics of GTFP and its driving mechanism were analyzed by applying global Malmquist-Luenberger index, non-parametric Kernel density estimation, spatial Markov chain and panel quantile regression. The results showed that: (1) During the study period, the evolution trend of GTFP was relatively stable, with fluctuation at the provincial and city levels, showing a "multi-peak" wave-shaped agglomeration pattern. (2) The spatial differentiation of GTFP had good continuity and stability. There was a possibility of leapfrog transfer in space, and the "Matthew effect" existed in the process of continuous transfer. (3) The level of economic development, industrial structure and population size all promoted the improvement of GTFP in the Qinghai-Tibet Plateau at different levels, while the acceleration of urbanization was not conducive to the growth of GTFP.