Abstract: The data validation and prediction are difficult yet important issues in the thermal vacuum test. In this paper, a typical solar panel thermal model is build, and the discrete equation is derived by the finite difference method. The coefficients are fitted by the least square method with “slip window” correction, to predict the temperatures of measuring points in the next measurement and control period. Correction slip window. Based on the test data of a certain solar panel, the prediction capability of this model and the related error analysis are made. In the end, the method to predict the heating power through the temperature of measurement points is discussed.