In order to improve the stiffness and reduce the mass of the manipulator with hybrid open and closed-loop kinematic chains, an optimization method based on nonlinear programming genetic algorithm (NPGA) is proposed. The stiffness mass ratio is selected as the objective function, the scale parameters are selected as design variables, the workspace, stiffness matching coefficients and the end stiffness are selected as constraint conditions, and the optimal structural parameters of the manipulator are obtained by using NPGA. The simulation results using finite element analysis (FEA) show that the stiffness of the manipulator is increased and the weight is reduced effectively after the optimization.