A new kind of piezoelectric sensor is presented and it can convert pulling forces into pressures without flexural moments. Thus it can be used to health monitoring of long-span structures with complex structural forms and high safety performance requirements. It makes full use of excellent properties of compression-resistance and fast sensoring of piezoelectric materials. A corresponding finite element model is established according to the sensors theory, and the modal equation is derived. Based on the maximum output principle, the objective function can be obtained, and then the optimization of the sensor arrangement solved by the genetic algorithm. In view of the uncertainty of the load and the randomness of the response signal, the wavelet analysis is used to process the collection of vibration signals, so as to avoid the observation overflow. Through simulation of intelligent monitoring on a long_span structure, results show that the designed senor with optimized arrangement can realize the monitoring effect on longspan structures. After the processing of the wavelet analysis, it can denoise while ensuring the stability of data process.