Gearbox is the most important component in automobiles. Gearbox can lead to serious damage if failure in any components occur and even lead to accidents. Thus the gearbox plays a vital role in a vehicle and hence condition monitoring is essential. Gearbox components state can be examined by having the vibration attributes which tells the gearbox state and condition. The main objective of this study is to propose a combinational classifier (Hybrid Classifier) for identifying and diagnosing faults in the gearbox using vibration as a solution to the problem. The vibration signals of the gearbox are acquired. Vibration Signals are used to extract the features by using the methods of feature extraction and selection of features are carried out before these variables are given as input to the classifier. Since no such procedures are obtainable to find productive number of attributes a detailed research is required to identify the best viable number of characteristics for given problem. The selected features are then classified using the proposed classification algorithm. The proposed classification accuracy of the technique has been compared with other machine learning approaches and discussed. The proposed fault diagnosis model can estimate the state and condition of process at any stage and prevents from unpredicted Failures. © 2017 IEEE.