The invention provides a forecasting method for a data safety model of a smart energy terminal. The forecasting method comprises the following steps: the step one, sensing the safe situation: according to the target of safety risk assessment, obtaining a data set collected by the smart energy terminal, completing the data fusion of the safety risk assessment to the data set, and obtaining abnormal situation data; the step two, analyzing the behavior characteristics: performing the situation index calculation to the sensed abnormal situation data, judging whether the index exceeds the predetermined safety threshold, once the index exceeds the predetermined safety threshold, identifying as an abnormal behavior element; and the step three: forecasting the safe situation: according to the identifying result of the abnormal behavior element, determining the safety risk countermeasure strategy and the safety margin adjustment strategy. The provided method is capable of performing the abnormal behavior analysis to the collected data set, and providing the data support for the next active defense.