Detecting defects in software at the bleeding edge of a software development life cycle is vital. Identifying defects before the deployment of software aids in delivering high-quality products, and reduces development costs. Machine learning techniques are deployed in the earlier stages of software development to improve software performance quality and decrease software maintenance costs. This study focuses on reviewing some papers published in software defect prediction using Machine learning techniques from 2020 to the current time to determine the predominance of machine learning methodologies adoption in software defect prediction. Google Scholar was used to source research papers for this study, and data was gathered from the publications. The process involves reviewing the selected papers, writing a concise synopsis of the papers, connecting and involving them where appropriate, reviewing existing methodology, and finally summarizing the findings. The result shows recent activities and trends in defect prediction research. This investigation will aid researchers in understanding the most recent and cutting-edge trends in software defect prediction research using machine learning techniques.