The indoor air temperature of greenhouse is an important parameter in environmental monitoring. Existing research on it has many, and the mathematical models have been made. But the obtained models do not apply to the northern areas features of dry and cold. On the basis of previous studies, the indoor air temperature model of greenhouse will be re-modeling by using the regression analysis method in this paper. Analysis each convection heat transfer coefficient of the model and re-established them. In the case of known the outdoor weather conditions, the solution of the indoor temperature will be implemented by using the computer programming. And the data of a period is selected randomly from the observational data to validate the model is feasible and applied.