This paper aims to obtain the optimal composite box-beam design for a helicopter rotor blade. The cross-sectional dimensions and the ply angles of the box beam are considered as design variables. The objective is to optimize the box beam to attain a target vector of stiffness values and maximum elastic coupling. The target vector is the optimal stiffness values of helicopter rotor blade obtained from a previous aeroelastic optimization study. The elastic couplings introduced by the box beam have beneficial effects on the aeroelastic stability of helicopter. The optimization problem is addressed by decomposing the optimization into two levels, a global level and a local level. The box-beam cross-sectional dimensions are optimized at the global level. The local-level optimization is a subproblem which finds optimal ply angles for each cross-sectional dimension considered in the global level. Real-coded genetic algorithm (RCGA) is used as the optimization tool in both the levels of optimization. Hybrid operators are developed for the RCGA, thereby enhancing the efficiency of the algorithm. Min-max method is used to scalarize the multiobjective functions used in this study. Optimal geometry and ply angles are obtained for composite box-beam designs with ply angle discretization of 10, 15, and 45°. © Springer-Verlag Berlin Heidelberg 2007.