Abstract: In order to realize the characteristics of temperature change in west of Inner Mongolia＇s greenhouse and build the greenhouse temperature prediction model, using hourly temperature observation data of inside and outside the greenhouse in town of Barunbieli in Alashan between January and April in 2013 to quantitative study the diurnal variation of temperature inside the greenhouse, then prediction model was established. The conclusion as following. First,the lowest temperature appears in 7：00 to 9：00 AM which is 6.4 ~10.9 ℃ and then fast rise 8.6 ℃ /h after 2 to 3 hour. The highest temperature appears in 14 to 15 PM which is 36.3~41.6 ℃ and then quickly fall 5.4 ℃/h. The temperature will change slowly after 19：00PM. Second, the greenhouses which have the conditions to know the real-time temperature can use the temperature fitting equation to calculate future temperature of each hour. Third, the method of regression analysis to established prediction model which including maximum temperature, minimum temperature and average temperature which can make good use of the temperature prediction results of next 3 to 7 days for farmers. Model has good degrees which can be used to predict relative changes of temperature. These results can provide theoretical basis for the development and utilization of the greenhouse climate resources and promote the development of agricultural facilities.