This paper proposes an automatic collision avoidance method to address safety challenges in near-ground flight for combat aircraft in uncertain environments. An adaptive controller is designed to address the escape maneuver control problem under gust disturbances and model uncertainties, combining tracking error feedback with disturbance estimation feedforward using a radial basis function neural network. The stability of the closed-loop system is proven through the Lyapunov method. To address the decision-making problem in aircraft escape maneuvers, a simplified six-degree-of-freedom model is established to reduce the computational load for trajectory prediction. Additionally, a multi-escape trajectory collision detection mechanism is designed to minimize maneuvering demands. Experiments show that the proposed controller achieves high accuracy and robustness, while the collision detection algorithm ensures fast computational efficiency and high safety. The algorithm also determines warning levels based on prediction time, enabling safe obstacle avoidance.