Building software that is free from errors is essential as software has become an essential component of human life. To plan an exceptional maintenance approach, software modules that are prone to defects must be identified before the software project is deployed. Early detection of software modules that are prone to defects can help create an appropriate process improvement plan in a time and cost-justified manner. This may also result in better product releases and high levels of client satisfaction down the road. Since defect prediction and measurement are indirect and dependent on multiple factors, they are critical issues in software development. Consequently, it would be more beneficial to choose a reasonable set of qualities that are pertinent and significant for defect prediction in any software module rather than taking into account all the attributes. Techniques for feature extraction and selection are used to identify the set of important attributes that are definitely influencing software module fault prediction Reducing the amount of characteristics following feature selection can improve the efficiency of prediction models, which can then be applied to identify faulty modules in a given set of inputs. The performance evaluation of several methods for feature extraction and selection is eventually drafted in this study. Four datasets from the PROMISE software engineering repository have been used in the experiment. The investigation and outcome indicate that using feature selection could increase accuracy.