Abstract: In order to seek the inside temperature prediction methods and indicators for sunlight greenhouse in Inner Mongolia, based on the temperature data inside and outside of sunlight greenhouse during 2011-2013, selecting 3representative stations from east to west, and the correlation between the outside temperatures and inside temperatures of the greenhouse were analyzed under the conditions of sunny day at different seasons（spring, autumn and winter）, different periods（day and night）. And on this basis, the prediction models of the greenhouse temperature daily variation, daily minimum and maximum temperature were established, then the accuracy of the models were fitted by 2014 observational data. Meanwhile the low temperature index of common fruits and vegetables in Inner Mongolia were collected and presented,and the outside temperature threshold according to the forecast models was inverted. The results showed that the models had goodness fitting accuracy, the inversion air temperature outside can provide the basis for decision-making, these results are of great worth for defending low-temperature disasters, it also can serve as references of temperature forecasting work compared for different regions sunlight greenhouse.