Brakes are indispensable element of automobile. It takes significant factor to slow down or stop vehicle at an instant which will help to prevent an incident or accident in panic scenario. In appropriate braking or breakdown in braking system may direct devastating effect on automobile as well as traveller safety. To enhance potential of braking system condition monitoring is drastic demand in automotive field. This research predominantly concentrates towards fault diagnosis of a hydraulic brake system with the principle of vibration signal. Feature extraction, feature selection and feature classification are the key measures under machine learning approach. Feature extraction can certainly accomplished by acquiring vibration from the system. Statistical features were for the fault diagnosis of hydraulic brake system. Best first tree algorithm will pick most effective features that will differentiate the fault conditions of the brake through given train samples. Fuzzy logic was selected as a classifier. In the present study, fuzzy classifier with the best first tree rules was used to perform the classification accuracy. © 2016. MechAero Foundation for Technical Research & Education Excellence.