Bearing fault diagnosis presents a great challenge to analysis as computational data. This analysis aims to propose a good fit approach to dealing with faulty data to optimize the correctness of data by developing a genetic algorithm (GA). Our result indicates the method of the genetic algorithm performs significantly better in several cases. The genetic algorithm performs with over 90% accuracy to overcome the faulty bearing fault data.