Abstract: Based on the CRU monthly mean global land surface data set of the University of East Anglia from 1980 to 2005, and combined with the Taylor diagnosis, the observation and model simulation of the surface temperature and its variation trend over the Tibetan Plateau were carried out. The annual average temperature variation trend is obtained from the historical data set of CRU observation data and the model, so as to predict the future surface temperature of the Plateau. The results show that most of the CMIP5 models have a good simulation effect on the annual mean temperature over the Plateau, but they are affected by seasonal differences. In addition, the average simulation ability of multi-model set is generally better than that of single model simulation, and it is closer to the data of observation field. The warming trend of the Qinghai-Tibet Plateau in the future is positively correlated with the intensity of external radiation forcing. The warming effect of the Qinghai-Tibet Plateau under the scenario of high emission is more intense than that under the scenario of medium and low emission, and the warming value is likely to exceed 6.5℃ by the end of the 21st century.