Considerable attention has been devoted recently to vibration control using intelligent materials as sensor/actuator. An intelligent control technique using a neural network is proposed for vibration control of micro-cantilever beam with bonded piezoelectric sensor and actuator. Structure modal characteristic analysis is done to determine the optimal configuration of piezoelectric sensor and actuator. With the piezoelectric elements are surface-bonded near the same position to the flxed end of micro-cantilever beam, an optimal controller, linear quadratic Gaussian (LQG), and an intelligent strategy based on neural network are investigated. Finally, the simulation results are given to demonstrate the effectiveness of the control method in this paper, neural network scheme has strong robustness to Gaussian disturbance compared with the LQG control method, Moreover, it is shown that vibration of micro-cantilever beam in MEMS can be significantly suppressed by the proposed neural network strategy.