Abstract: In view of the limitations of the traditional reentry trajectory optimization method, which has slow convergence speed and high sensitivity to initial values, a fast algorithm for reentry trajectory based on sequential convex optimization is proposed. The method takes the change rate of inclination angle as the control variable which improves existing convex strategy and uses B-spline curve to discretize the control variable, which can effectively suppress the sawtooth phenomenon in the process of numerical optimization due to the numerical discrete method. At the same time, to avoid the problem that the algorithm appears pseudo-infeasible near the initial guess value, additional virtual control and a "backtracking straight line" search method are added to improve the stability of the algorithm, these can improve the stability, rapidity and smoothness of the algorithm. In order to study the aerodynamic parameter perturbation in the reentry process of aircraft, a generalized polynomial chaotic theory research method with few sampling points, easy implementation and high computational efficiency is proposed. A robust optimization model of reentry trajectory based on the combination of generalized polynomial chaos and convex optimization is established. The model considers the effect of aerodynamic parameter disturbance on the optimization results during the optimization process, which can avoid the complex iterative design of conventional trajectory and guidance law. What's more, it can effectively reduce the sensitivity of optimal trajectory to aerodynamic parameter perturbation, and under the disturbance of uncertain aerodynamic parameters, which still can ensure the safety of the aircraft smooth completion of the mission. At the end,taking the reentry mission of a reusable aircraft in the United States as an example, the rapidity of reentry trajectory optimization method based on sequence convex optimization and the anti-interference ability of robust optimization model to aerodynamic parameter perturbation are verified and it shows that this method has certain engineering application.