This paper proposes an approximate method based on hierarchical BP neural network and Bezier curve fitting to efficiently solve the reentry domain of aircraft. Based on the trajectory planning method, this approach utilizes the maximum longitudinal point and minimum longitudinal adjacent points, as well as the maximum and minimum transverse points of the reachable domain as nodes. It employs Bezier curves to approximate and replace the boundary of the reachable domain, transforming the problem into solving special points of the reachable domain boundary and control points of the Bezier curve. The initial flight states and Bezier curve control points corresponding to each other are utilized as the input and output of the neural network, and a training database is established. The neural network model is trained hierarchically, resulting in a neural network model that can efficiently solve the reachable region. Simulation results demonstrate that compared with existing algorithms, this proposed method achieves rapid solution under different initial input states while meeting online calculation requirements with improved precision.