Abstract: Change in phenology is among the first signals of adjustments in species responses to climate anomalies and is of great significance to clarify the interference mechanism of extreme temperature on grassland phenology. In Qinghai-Tibet Plateau, plant development is commonly regulated by both temperature and precipitation. To better understand how grassland phenology may respond to extreme temperature events, the spatial-temporal dynamics of start of growing season (SOS) and the sensitivity to different extreme temperature events of different grassland were explored based on GIMMS NDVI, grassland types and extreme temperature data on the Qinghai Tibet Plateau from 1986 to 2015. In this study, with the assistance of Sen gradient, variation coefficient, Hurst index, geographical detector and partial correlation analysis, the spatio-temporal patterns of the start of the growing season (SOS) in the grassland were investigated by using the Normalized Difference Vegetation Index (NDVI) on the Qinghai-Tibet Plateau (QTP) from 1986 to 2015. At the same time, the extreme temperature regulating grassland spring phenology under the ongoing conditions of climate change was analyzed. The results show that: (1) the SOS concentrated in the period of 120-130 days of the year. The SOS appeared first in plain grassland, meadow and alpine and subalpine meadow, basically before the 130th day. The SOS of desert grassland, alpine and subalpine plain grassland and slope meadow appeared late, concentrated in the period of 125-145 days. (2) The SOS showed a significant advancing tendency in all types, and the rate concentrated at 0~±1.5 d·a^-1. The predicted change trend of SOS in the coming period of time is opposite to the last change trend. (3) SOS of meadow was the most sensitive to extreme temperature, followed by plain grassland, alpine and subalpine meadow, alpine and subalpine plain grassland, and desert grassland was the least sensitive to extreme temperature. (4) The correlation between cooling related extreme temperature and grassland SOS is mainly weakly positive correlation, and the correlation between warming related extreme temperature and grassland SOS is mainly weakly negative correlation, and extreme high negative correlation exists especially in the Himalayas, Kunlun Mountains, Tianshan Mountains and Hengduan Mountains with high altitude. The results of this study could provide new understanding on the impacting factors to the phenology of grassland of Qinghai-Tibet Plateau, and could also provide effective measures for grassland disaster prevention and remediation, for grass seeds breeding, grassland development monitoring, climate anomalies responding and vegetation ecological restoring.