Abstract: The precooler is an important component for air cooling in engines, and it can be costly to conduct numerical simulations on the precooler directly. Developing accurate and rapid prediction methods are crucial for the design and condition monitoring of such engines. In this study, based on the numerical simulation of flow and heat transfer in the microtubes of the precooler, a rapid prediction model for flow and heat transfer was established using the back pressure of the precooler outlet as a variable, employing the proper orthogonal decomposition theory and the radial basis function interpolation method. After that, rapid prediction of non-sample operating conditions was conducted. The research shows that the rapid prediction method can obtain the flow and heat transfer information for any non-sample operating condition in less than 1 s and achieve sufficiently high flow field reconstruction/prediction accuracy using only the first 4 modes. Predictions of flow fields such as Mach number, temperature, and total pressure are in good agreement with the numerical simulation results, with a maximum prediction error of no more than 2.3% and the prediction error of outlet total pressure recovery coefficient not exceed 4.8%, confirming the accuracy of the present method. The study also shows that when the pressure of the precooler outlet is very low, there will be a local supersonic zone at the tube outlet, and the prediction error is mainly concentrated in this area. The appearance of the supersonic zone will cause a significant total pressure loss, thus this phenomenon should be avoided as much as possible in the design or operation of the precooler.