The invention discloses a MA-DBN-based GIS decomposed gas content prediction method, which comprises the following steps of firstly, acquiring the SO2, SO2F2, SO2F2 / SO2, air temperature, humidity andaltitude data in a period of time monitored by a state grid for the GIS equipment; secondly, constructing an MA model, taking the collected gas data as the input of the model, and predicting the gas data in the future time period; and finally, constructing a DBN model, taking the predicted gas data as the input of the DBN model, and correcting the DBN model to obtain the accurately predicted gas data in the future time period. The method solves the problem of inaccurate MA algorithm model prediction, has the advantages of being high in convergence speed and not prone to falling into local optimum, can eliminate the fault before the fault of a GIS switch device happens, and reduces the economic losses caused by the faults of the GIS equipment.