The friction vibration signal corresponding to different fault types of brakes is significantly different, thus a hydraulic disc brake fault diagnosis and judgment scheme based on vibration signal detection was proposed and validated in the paper. According to the composition and electric control principle of hydraulic brake, the vibration signal detection and brake pressure control strategies were established. The key hardware and isolation circuit of the detection system were designed, which can effectively improve the stability and anti-interference ability in the process of vibration data acquisition and transmission with SPI and CAN bus communication scheme. In order to eliminate noise interference in friction vibration signal and improve the accuracy of fault diagnosis with advantage, wavelet threshold denoising method was applied to brake disc vibration signal processing. The denoised signal can be obtained by inverting the selected wavelet coefficients to separate the useful components of the signal from the invalid noise. For the research on the correlation between fault types and vibration signal characteristics, a friction vibration test bench for hydraulic braking system was built to obtain and process the output signal from vibration sensor. The frequency spectrum response laws of vibration signals were compared and analyzed when the brake disc was normal and had surface defects, including microcracks, excessive wear and fatigue peeling. The external parameters such as brake pressure and spindle speed were keeping constant, thus eliminating the influence of working factors on vibration characteristics. In order to study the influence of brake pressure failure on friction vibration, the variation law of vibration signal is compared and analyzed under the condition of hydraulic valve leakage and pressure maintaining failure. The results show that the friction vibration signal can be used as an effective basis for judging the surface defects and brake pressure faults of brake discs. Under the condition of wavelet threshold denoising, accurate and effective vibration parameters can be obtained, which is of great significance to the safety and reliability of the braking system.