The ascent trajectory design for Rocket-Based Combined Cycle （RBCC） hypersonic vehicle has many typical characteristics， including complex power system working modes， strong coupling between thrust and flight state， highly nonlinear models， numerous complex constraints， parameter uncertainties， etc. In this paper， a robust trajectory optimization method for RBCC power ascent based on non-intrusive polynomial chaos， the Gaussian quadrature strategy， and sequential convex optimization is proposed to enhance the trajectory’s anti-interference ability and process reliability. Firstly， a robust trajectory optimization model for the RBCC hypersonic ascent， accounting for parameter uncertainties， is established. An uncertainty quantification propagation algorithm based on the Gaussian quadrature strategy and non-intrusive polynomial chaos is designed to transform the robust optimization model into a deterministic trajectory optimization problem with extended dimensions. Then， the extend model is convexified and discretized using convex optimization theory， and a trajectory optimization solution strategy based on sequential convex optimization algorithm is established to achieve a solution of this high-dimensional deterministic optimization problem. The optimaization results of a certain air-based vehicle’s ascent trajectory indicate that based on the constructed model and trajectory method， the robust trajectory optimization of the ascent phase of the RBCC hypersonic aircraft can be effectively completed， and the optimization results are in line with the working characteristics of the RBCC power system. Compared with the traditional deterministic trajectory optimization algorithm， the proposed method can effectively reduce the influence of random disturbances on the ascent trajectory， thereby improving the reliability and robustness of the trajectory.