Warming-induced shifts in vegetation phenology affect ecosystem functions and the hydrological cycle. However, research has so far focuses on the trends and driving factors of phenological changes in temperate and tropical forests, while little is known about changes in grassland phenology in arid and semi-arid regions and their effects on gross primary productivity (GPP) of ecosystems. Therefore, an in-depth study of plant phenology and its relationship with productivity in grasslands under a warming context is essential to predict the response of grassland ecosystems to future climate change and the regional carbon cycle. Based on the meteorological data and GIMMS NDVI data from 1982-2015, this paper analyzed the changes in the start of the growing season (SGS) and end of the growing season (EGS) of temperate grassland vegetation in China and their responses to climate change. We also quantified the contribution of plant phenology to the dynamic change of GPP with the help of first-order difference method. The results showed that (1) in general, the increasing in nighttime temperature (T_min) in the 1-2 month preseason phase significantly advanced SGS, however, the daytime temperature (T_max) from the current month to first two months before the SGS had a weak effect to advance the SGS. Additionally, cumulative precipitation in the first three months before the SGS had the strongest effects on the advancement of SGS. Cumulative solar radiation was found to have a relatively weak effect on SGS in all preseason periods. (2) In terms of the EGS, the results showed that different preseason time scales of diurnal temperature showed an opposite effect on the EGS in temperate grassland, and higher nighttime temperatures would delay the onset of EGS. Short-term cumulative precipitation had a significant effect on the delay of the EGS and reached its strongest in the first 2 months of the season. The delayed effect of cumulative solar radiation on the EGS gradually changed to an advancement effect with the increase of preseason scale. (3) Based on the results of the relationship between phenology and GPP, both SGS and EGS contributed positively to the inter-annual variation of GPP in grassland across the region, and the relative contribution of EGS was stronger than that of the SGS. The results of the study help deepen the understanding of the interactions between terrestrial ecosystems and climate change and carbon cycle, and provide scientific basis for grassland adaptation to future climate change and ecological construction.