Satellite-derived products and field measurements verify that Tibetan Plateau (TP) has been experiencing continuous vegetation greening and productivity increase; however, it remains unclear how this greening translates into productivity and how long-term productivity variations depend on greenness across the TP. Moreover, ignoring the accuracy evaluation of satellite-derived greenness and productivity products may mislead the understanding of TP vegetation changes. Thus, we initially assessed the accuracy of three widely used leaf area index (LAI, proxy of greenness) products (i.e., MODIS, GLASS and GEOV2 LAI) and three gross primary productivity (GPP, proxy of productivity) products (i.e., MODIS, GLASS and PML-V2 GPP) to selected robust products to represent greenness and productivity respectively. Then, we explored the dependence of spatiotemporal GPP dynamics on greenness variations during 2000-2020. Results indicated that PML-V2 GPP and MODIS LAI were more robust and reliable than other satellite-derived products when compared to the reference values. They revealed a prevailing increase in GPP over the past two decades, with a regional average of 71 % higher than that of LAI. Notably, the area proportion of significant productivity enhancement was 31.6 % higher than that of significant greening. About 24.7 % of the TP displayed significantly inconsistent trends. The dependency of GPP on LAI gradually decreased with the increasing water availability, the complexity of vegetation structures, and dense canopy community. By calculating leaf photosynthetic capacity, we found that this indicator greatly regulated the velocity discrepancy between GPP and LAI, and the contribution of only greening to productivity is limited, only occupying 11.9 % of the TP, which was helpful in understanding the inter-annual changes of vegetation dynamics under varying environment conditions. We therefore reveal an unexpected rapid increase in productivity than greening during 2000-2020 on the TP, as well as highlight the caution of only using satellite-derived greenness indicators for assessing long-term changes in vegetation productivity dynamics, especially over mesic ecosystems with complex vegetation structures and dense canopies of TP.