Abstract: To improve facility agro-meteorological services in Jinshan Shanghai, and mitigate the impact of meteorological disasters on greenhouse ‘Yulu’ peach production, based on the meteorological observation data inside and outside the solar greenhouse in spring during 2015 to 2017, the temperature characteristic and the minimum temperature forecast model inside the greenhouse in spring were studied by using correlation and stepwise regression analysis method. The results showed that the temperature inside the peach greenhouse had an obvious diurnal variation in sunny and cloudy days, and the range of temperature change was greater in sunny days than in cloudy days. The humidity inside was higher than that in continuous overcast days, and it seriously affected the normal growth and pollination of flat peach. The minimum temperature inside the solar greenhouse was preferably related to the highest and lowest temperature inside or outside the solar greenhouse. The minimum temperature forecast model inside the solar greenhouse in winter has been set up by using 2015—2016 spring data. The forecast model has been verified by the spring data in 2017. The absolute error (AE) of the forecast minimum temperature inside the peach greenhouse in the whole spring was about 1.56℃, the average relative errors was about 18%, and the average absolute errors (RMSE) was about 4.97℃,The model result in overcast days was best. The results of model fitting and forecast test showed that forecast model had high accurate and good practicability.