The present smart composite structures are composed of PZT actuators and graphlte-epoxy composites to improve the performance of vibration suppression. The control performance of PZT actuators strongly depends on their placements and vibration characteristics of laminated composites are varied by their lay-up configurations. Thus, a multidisciplinary optimization method employing actuator placements, lay-ups of laminated composite plates and control systems as design variables is proposed here. First, a database containing the frequencies and modal matrixes is made to avoid the FEA repeatedly. Then, the H_2 norm of control response is minimized with respect to the above design variables with referring the database by a simple genetic algorithm with assuming that the state feedback is available. Finally, the output feedback system is reconstructed with the dynamic compensator based on the linear matrix inequality (LMI) approach. The effectiveness of the present optimization technique is confirmed by the experimental and numerical results.