Software maintainability and reliability are important components in making sure that the software systems can utilize for a long time. Process metrics in conjunction with machine learning approaches used to enhance software maintainability and reliability. The supervised machine learning approach predicts the notion that automated processes supported by machine learning algorithms SVM, DT, and NB can assist in more precisely identifying software faults. A technique for utilizing data and analytics to anticipate, assess, and improve software system performance is the use of process metrics and machine learning to enhance software maintainability and reliability. Use some of the attributes of code quality. The problem is reliability and maintainability of software systems are still use of conventional software development methods, which increase costs, extend development durations, and slow time-to-market. The major goal of this study is to use process measurements and machine learning approaches to improve software maintainability and reliability. Our data was collected from open-source and by reading different articles. Software maintainability dataset from figshare.com and Github. Machine learning techniques Support Vector Machine, decision tree, and Naive Bayes to classify software bugs or defects based on various metrics from a combined dataset of software maintainability and reliability data such as Combine (maintainability + reliability), vg, and LoCodeAndComment. The results suggest that utilizing a combination of process metrics and machine learning can significantly improve the performance of software systems, reduce maintenance costs, and improve overall maintainability and reliability. Combine (maintainability + reliability), vg, and LoCodeAndComment parameters used we take the most accurate enhancement model: 98.4% in SVM, 97.4% in NB, and 98.4% for DT machine learning algorithms. Keyword Process metrics, software maintainability, software reliability, PROMIS, machine learning