During high angle of attack flight, high-speed projectiles exhibit phenomena such as intense flow separation, shock-shock interference, and shock-boundary layer interaction, rendering their aerodynamics highly nonlinear[1]. Traditional linear models fail to accurately characterize the strong nonlinear aerodynamic forces acting on high-speed projectiles, whereas black-box modeling methods like deep neural networks, despite possessing strong nonlinear fitting capabilities, suffer from poor interpretability and are prone to overfitting. To address these limitations, this study proposes a Symbolic Regression Aerodynamic Modeling Method Considering Physical Constraints. This method incorporates the energy conservation equation of high-speed projectiles as a physical constraint into the loss function of aerodynamic modeling. Leveraging the powerful nonlinear fitting capability of symbolic regression[3], it is possible to obtain an accurate nonlinear aerodynamic model with an explicit expression. The method is applied to a publicly available ARF(Aeroballistic Research Facility) high-speed projectile. The training set consists of trajectory data with an initial pitch angle of 30 degrees, and the test set consists of trajectory data with an initial pitch angle of 15 degrees. The results demonstrate that this method can accurately identify the nonlinear aerodynamic damping of high-speed projectiles using only a single trajectory, exhibiting excellent generalization performance on the test set.