This paper proposes a multidisciplinary design optimization methodology for a passive morphing composite Unmanned Aerial Vehicle (UAV) wing utilizing surrogate modeling and evolutionary algorithms. In the present research, minimization of the drag coefficient (CD), structural mass (m), and root bending moment (Mroot) was pursued together with favorable aeroelastic twisting properties. A neural network-based surrogate modeling technique was applied in order to precisely evaluate the aerodynamic/structural responses as an alternative to an expensive high-fidelity simulation process. The surrogate model was embedded in the evolutionary multi-objective optimization process, whereby the decision variables were determined by the composite laminate stack up to take advantage of the passive aeroelastic twisting of the wing structure. The obtained results have shown very good prediction capability of the surrogate model, with drag coefficient error lower than ±0.002 and aeroelastic twisting error ±0.5° with respect to the validation data. Pareto-optimal solutions exhibit well-balanced solution sets among the considered conflicting criteria, with the majority of optimal designs located at CD ≈ 0.030-0.034, structural mass = 4.3-4.8 kg, and bending moment of 240-300 N.m. Twist values ranged from −2° to 8°, allowing passive morphing. Results indicate the promise of passive morphing in lightweight UAV systems. The approach offers an efficient computational tool for navigating through intricate aero-structural design domains.