Abstract: The paper targets at promoting the development of agricultural facilities, expanding the scope of agro-meteorological services, and reducing impact of chilling damage on greenhouse agricultural production. It made researches on the temperature variation in the solar greenhouse and established related predictive models, in accordance with meteorological observation data inside and outside solar greenhouse collected on multiple monitoring sites over the years, by adopting correlation methods and the temperature variation rate analysis methods. It was shown from the results that there appeared significant diurnal and seasonal changes in temperature of solar greenhouse. At different weather conditions and different seasons, the variation rate varied; the temperature variation rate in sunny days was greater than that in cloudy days; the temperature variation rate in cold period was greater than that in non-cold period; the diurnal temperature variation of solar greenhouse was characterized by “one-increasing-factor” and “two-decreasing-factor”, seasonal change presented a “U-shaped” curve. The inside/outside temperature difference of solar greenhouse increased with decline of outside temperature, showing a parabolic change from December to February next year. The greenhouse temperature prediction model was applied to predicting minimum temperature inside greenhouse in February 2012. The average absolute error of forecast value was -0.7℃, and the average relative error was -11.4%.