Abstarct: Abstract Nowadays, the development of reliable and high-quality software is one of the main challenges discussed in software engineering. That's why we face the issue of quality assurance, which is a time-consuming process and requires a lot of resources. Software bug prediction can be a more cost-effective and optimal method by allocating more limited resources, and it can be used by predicting the susceptibility of software modules to bugs before testing. One of the challenges in the field of software defect prediction is that the previous prediction models were old models or used data sets that were small, limited and old projects; Due to the changing process of software development, these models do not provide much functionality for today's software. In this research, a new dataset called BugHunter has been used to help the process of bug prediction compared to old methods. This valuable and relatively large dataset consists of 15 popular open source Java programs that are widely used by software engineers today. Various machine learning techniques and algorithms have been used to date in the field of software bug prediction. The challenge in these techniques is that the accuracy of existing prediction models is low for this data set. This research aims to increase the performance of shape prediction model by combining clustering and classification techniques to increase the accuracy of the prediction model. Among the advantages of the presented model, we can mention the high accuracy of the presented model as well as less processing than other forecasting models. The f-measure obtained from the implementation of the presented model on the Bughunter dataset was 88%. The results are an average of 7% better than the best method ever seen.