Abstract: To improve aerodynamic performance and stealth performance of aircraft synthetically, in this paper, a gradient optimization method based on decomposition was used in the aerodynamic and stealth integrated optimization design of a transonic wing. Based on Tchebycheff approach, the multi-disciplinary multi-objective optimization problem of aerodynamics and stealth was decomposed into several single-objective optimization subproblems. Each single-objective subproblem was solved by using gradient optimization methods. The gradient of the aerodynamic objective parameters to design variables could be obtained via the discrete adjoint equations. The physical optics (PO) code is differentiated through the automatic differentiation, thus the gradient of the radar cross section (RCS) to design variables would be obtained. Every branch solution corresponding to the weight coefficient given in the optimization set is obtained by integrated optimization design. Both the drag coefficient of the optimized wing and the mean value of RCS at key position decrease, and the lift-drag ration of the optimized wing increases, as compared with the original wing. It is verified that the optimization method has better practicability.