Since it was difficult to predict the pressure drop of missile engine, according to the measured data of engine pressure of a certain type of missile, the measured data were interpolated, then the five sets of data were used as input data, the latter set of data as forecast data, and the model identification of missile engine pressure was carried out by using support vector machine ensemble method to realise the prediction of missile engine pressure. Through the simulation analysis of three groups of engine pressure measured data, it was found that the maximum forcast error of support vector machine ensemble was 0.102%, which met the requirements of missile engine pressure prediction, and it played an important role in preventive maintenance of missile engine pressure.