This research explores the application of Predictive Learning Analytics (PLA) in evaluating and forecasting student performance and engagement on Moodle. By analyzing student activities such as logins, assignment submissions, quiz scores, and forum interactions, the study employs machine learning algorithms to generate predictive insights that help educators support students proactively. A range of models, including Decision Trees, Random Forests, Support Vector Machines, Logistic Regression, and K-Nearest Neighbors, were applied, with the Random Forest model demonstrating the highest accuracy and predictive capability. Results suggest that PLA can accurately identify students at risk of underperforming, enabling timely intervention and personalized support. This data-driven framework fosters adaptive and responsive educational environments, enhancing student engagement and retention by tailoring resources and interventions to individual needs. The study underscores PLA’s transformative potential in digital education and contributes to the growing field of predictive analytics in educational settings, offering a valuable tool for optimizing student success.