Abstract:The energy consumption forecast of airline directly affects energy demand planning and energy saving decisions. A single prediction model is not sufficient to fully describe the trend of airline’s energy consumption, which leads to the prediction accuracy is not high. Considering the trend and volatility of the data, a combined sliding prediction model of the energy consumption based on the metabolic Gray Markov-ARMA is put forward to improve the prediction accuracy in this paper. Firstly, the Gray Markov is used to describe the trend of energy consumption. Secondly, ARMA model is used to describe the volatility by exploring the correlation between the residual sequences. Thirdly, Metabolic is used to overcome the timeliness loss of old data. The simulation results show that the prediction accuracy of the metabolic Gray Markov-ARMA model is higher than that of the ARMA model and the Gray Markov model, so it can effectively predict the monthly energy consumption. The study of this method can provide support for the monitoring of energy consumption of airlines and the optimization of energy saving work.