Abstract: The Integrated Biosphere Simulator (IBIS) is used to evaluate the spatial and temporal patterns of annual and monthly top 1 m soil temperature across China for the period 1955-2006. Observed soil temperature data collected from 650 national meteorological stations were used to validate the simulated soil temperature in spatial and temporal scales. Mean error (ME), root mean square error (RMSE) and coefficient of determination were used as the performance criterion. The validation results presented that IBIS model performed better in the southern China than in the northern China and the Tibetan plateau region. The mean error and root mean square error are less than 2℃ in the southern China and more than 3℃ in some area of the northern China and the Tibetan plateau region for annual soil temperature. Monthly validation results showed that the model performed better in spring, summer and autumn than in winter. The mean error was bigger than 3℃ in winter and it was between -1.4℃ and 2℃ in other three seasons. The root mean square error was larger than 3.5℃ in winter and it was between 2℃ and 3℃ in other three seasons. The root mean square error and coefficient of determination also indicated better performance in the southern China than in the northern China and the Tibetan plateau region for monthly soil temperature simulation. The coefficient of determination showed that the model captured the spatial variance of soil temperature well. Based on the simulated soil temperature, trend analysis was applied for annual and monthly soil temperature from 1955 to 2006 using Mann-Kendall method. Annual soil temperature presented significant increasing trends in the northern China and slight increasing trends in the southern China. Small areas of Sichun Basin, center part of Guizhou Province, southeast part of Tibet, and Tianshan Mountains region presented significant or slight decreasing trends for soil temperature. Monthly soil temperature trends analysis presented significant increasing trends in most areas of the northern China and in the southern China the spatial pattern of trends for each month was different. From July to September, the soil temperature presented decreasing trends in most area of the southern China. In August, the soil temperature presented significant decreasing trends in the southern China. Uncertainty should be considered in soil temperature simulation. Uncertainty in observed data used to validate the simulated results, in input data such as soil texture, atmospheric boundary condition including solar radiation, air temperature and precipitation required by the models, and in spatial scale matching will make the soil temperature estimation diverge from the true state because of gradually errors accumulating.