Abstract: The study of the response of grassland vegetation phenology to climate change is the key to understand the history of regional land ecosystem productivity and the mechanism of future change. Based on MODIS NDVI data, grassland type data and ground measurement data, combined with en trend analysis and partial correlation analysis, the temporal and spatial characteristics, trends, stability of changes and their variation characteristics with climate factors of grassland vegetation phenology 2001-2020 were studied. The results show that: (1) with the increase of altitude, the start of growing season (SOS), end of growing season (EOS)and length of growing season (LOS) changed irregularly. (2) SOS was concentrated in mid April to late May (110-150 d), and 79.32% of the area showed a significant advance trend, with a rate of 0-1.5 d·y^-1, the SOS of mountain meadow < temperate grassland, shrub grass and desert < alpine < alpine grassland. EOS concentrated in mid September to mid October (260-290 d), show a delayed trend. The LOA lasted for 130-180 d, show a prolonged trend. (3) The decrease of annual precipitation, the increase of spring temperature, the increase of extreme climate index of temperature and the decrease of maximum 5 d rainfall will lead to the advance of grassland SOS. The increase of annual average temperature, maximum 1 d rainfall, summer and winter temperature, and the decrease of autumn temperature made trend of grassland SOS. (4) The increase of annual precipitation, daily maximum temperature and autumn temperature delay the grassland EOS, while the increase of annual average temperature lead to the advance of grassland EOS. The above results can provide a new understanding of the influencing factors of grassland phenology in Qinghai Province, and also provide a reference for grassland to adapt to future climate change and grassland breeding.