A constrained multiobjective optimization code, based on genetic algorithms, was applied to improve the aerodynamic performance of a commercial hang-glider wing in some points of the hodograph starting from the shape of an existing vehicle. In particular, the optimization process was aimed to maximize the efficiency at trim speed and to minimize the stall speed. Constraints were imposed on the desired lift and on the characteristics of the longitudinal static stability. Results, in terms of Pareto front approximation and in terms of improvements of the performance of the base wing, are presented and discussed. Even if the work has to be considered only the first step of a totally automated, numerical design approach of a hang-glider wing, obtained results demonstrate how few hours of computation can successfully substitute years of trials and errors.