Identification and removal of software defects is tedious and time consuming for software development. Improperly planned projects could have defects and the time spent to spot and fix them requires more than the code development time. A reverse engineering sub-area is identification of modules necessitating re-engineering; focusing on faulty modules prediction based on existing information sources like documentation and source code. Predicting defective module is essential in maintenance and reuse by simplifying system working with information and reusable parts localization. In software defect prediction, predictive models estimation is based on code attributes to assess software modules containing errors likelihood. In this paper, the classification accuracy of Boosting techniques for software defect prediction based on the KC1 dataset is investigated.