In software engineering, ensuring high software quality is paramount for successful project delivery and user satisfaction. This study presents a robust approach to software quality prediction utilizing machine learning techniques combined with source code metrics. Five models were rigorously evaluated: Random Forest, Gradient Boosting, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and a novel Hybrid Model that integrates the strengths of these algorithms. The models were assessed using a comprehensive dataset sourced from the Software Metrics Data Repository, specifically focusing on various performance metrics including accuracy, precision, recall, F1-score, and AUC-ROC. The results revealed that the Hybrid Model outperformed its counterparts, achieving an accuracy of 93%, precision of 91%, recall of 90%, an F1-score of 91%, and an AUC-ROC of 0.94. These findings underscore the efficacy of the Hybrid Model in improving software quality prediction, presenting a significant advancement in the field. The proposed methodology not only enhances the reliability of quality assessments but also sets a precedent for future research in software quality engineering, paving the way for more sophisticated, data-driven decision-making processes. In conclusion, the proposed framework offers a scalable, accurate, and efficient solution for defect prediction and maintainability assessment, addressing critical challenges in software engineering. Its superior performance and adaptability highlight its potential for practical application in real-world software development and maintenance processes.