The vibration signal is very important to the bridge healthy monitoring, because the variation of bridge vibration frequency indicates the changing of bridge structure state. Wavelet analysis theory is applied in this paper and study to analyses the monitoring data of bridge vibration signals in the bridge health monitoring system, and through multi-scale decomposition and reconstruction, as well as wavelet denoising, to enable classification and feature extraction from the vibration signals. And meantime through comparing with fast Fourier transform (FFT), it has demonstrated the validity of taking wavelet analysis as an instrument to analyses the vibration signal of bridge structure health monitoring.