A project's source code change history contains the modification that generates a defect and the change that fixes it. This defect-generation and fix experience can be used to forecast future defects. This research paper presents defect forecasting algorithms that analyze a software project's change history. Defects are related. Our algorithm finds this relation by storing locations that are likely to have defects. That is useful to find most defect prone files. An evaluation of open source projects with more than 5,000 revisions shows that the selected defect training data account for 72%-90% of future defects .By leveraging project history and learning the unique defect patterns of each project, both approaches is used to find locations of defects. This information can be used to increase software quality and reduce software development cost