Uncertainties in local solar radiation, ambient temperature and thermal load data have been one of the major factors limiting the reliability and efficiency of solar thermal hybrid systems. In the present paper, moving average auto regressive exogenous (ARX) model based reasoning has been mooted and modified to include moving average method, as an effective tool for predictions of these data. The results show that the method is quite robust and is capable of predicting fairly accurate results, which would make these systems more viable in areas where meteorological data are not available or vague.