Predicting software defects has become a significant area of research in the field of software engineering. By concentrating on faultaprone modules, the precise prediction of defectaprone software modules can aid in software testing efforts, lower expenses, and enhance the software testing process. Data sets on software defects are asymmetrical, with relatively few faulty modules in comparison to those without defects. The presence of noisy attributes in the dataset leads to a considerable fall in the performance of software defect prediction. In this study, we suggest combining the bagging technique with evolutionary algorithms to enhance software defect prediction performance. To address the issue of feature selection, a genetic algorithm is utilised, while the bagging technique is utilised to address the issue of class imbalance. The range of applications for genetic algorithms is expanding quickly, according to a survey of papers on the subject. The primary objective of the writers is to offer a productive feature selection for the advancement of the study