Ram air turbine (RAT) system is the emergency power generation system of the aircraft. It supplies power to the aircraft after the aircraft loses the main AC power. RAT is stowed in the RAT bay in the non air tight area normally and deployed into the air flow under emergency conditions. The temperature of the RAT bay determines the deploy time of the RAT, thus affecting the safety of the aircraft under emergency conditions. The temperature prediction of the RAT bay is of significant for the safety under emergency condition. In this paper, the RAT bay temperature, altitude, atmospheric static temperature, flight speed, and cabin temperatures around the RAT bay are analyzed as a whole through the Proper Orthogonal Decomposition (POD), and the RAT bay temperature is predicted under specific working conditions through RAT bay temperature model and the mapping relationship between the simplified model and the complete model. This paper selects the test data as the sample data, and uses POD method to predict the temperature of the RAT bay. According to the prediction results, the energy distribution of the first order mode exceeds 99% of energy of all modes. After data normalization, the absolute average error of the prediction results could reach 0.063 as the minimum, and the variance could reach 0.07 as the minimum; furthermore, the resuls show that the better the prediction effect could be achieved with more samples.