This paper proposes a binary particle swarm optimization (BPSO) and binary Genetic Algorithm (BGA) in feature selection process using different fitness functions in the field of bearing fault diagnosis. The vibration data obtained by extracting vibrational signals, considering four cases such as Normal Bearing, Inner race fault, Outer race fault and Ball fault, at constant speed conditions. This paper proposes four fitness functions applied on BPSO and BGA that gives rise to a reduce feature set. Furthermore, standard classification algorithms such as KStar, Naïve Bayes, JRip, J48 are used. Furthermore, Neural Network classification algorithms such as Back Propagation Network (BPN), RBF network and Deep Neural network (DNN) apart from standard classifiers are used. The aim of the paper to identify an appropriate combination scheme contains feature selection method based on appropriate fitness function algorithm along with better classifier that maintains high accuracy with adequate computational time. It was observed that GA possess higher accuracy than BPSO, yet the computation time is high. It is observed that PSO based schema of feature selection are optimal with high levels of classification accuracy and optimum computational time.