Gears are a vital element considering its applications in a variety of machine tool applications. An unpredicted failure of the gear may cause substantial economic losses. For this very reason, fault diagnosis in gears has been the subject of intensive research. An intelligent method to diagnose and predict the gear fault using vibration signal is proposed in this research work. A signal which contains useful information from various conditions of gear is extracted from an experimental rig. From the original acceleration vibration signals, statistical features are extracted after using the EMD (Empirical mode decomposition) which decomposes the signal into a finite number of stationary intrinsic mode functions (IMFs). Extracted features are recognized and classified by a novel heuristic classifier artificial bee colony (ABC) algorithm. In order to select the predominant features, traditional Genetic algorithm (GA) as well as ReliefF Genetic Algorithm (RFGA) is utilized. The fault diagnosis results are compared with support vector machine (SVM) classifier and their relative efficiency were compared based on the classification accuracy. © February 2015 IJENS.