Abstract: The deep integration of wind tunnel experiments, numerical calculations, and model flight tests is an inevitable requirement for developing future high-speed vehicle. This article focuses on constructing high-speed wind tunnel facilities, numerical calculation software and hardware, model flight testing capabilities, and self-developed flight test techniques for measuring the surface temperature, heat flux, pressure, friction drag, etc. Based on the characteristics of aerodynamic data fusion, an aerodynamic data fusion method, based on the correlation between aerodynamic data and physical models as well as deep learning, was developed to solve the problem of data association between different data sources, significantly improving the data's reliability. The data fusion method has been successfully applied in the flight-ground correlation of the pitching moment coefficient and aerothermodynamic data for high-speed vehicle. A comprehensive study of the high-speed shock/boundary-layer interaction problem using various methods results in an empirical formula of the interaction force/thermal load, a correction to the pressure/heat flow correlation, and the verification of the low-frequency oscillation of the separation bubble under real flight conditions.