This paper explores the application of learning analytics in predicting students' performance (SP) within Moodle, a widely used Learning Management System (LMS). The study focuses on measurable academic progress and outcomes, aiming to assist educators in early identification and resolution of issues to boost student productivity and success. Our approach began with a literature review to identify predictive attributes for student performance. We then collected and analyzed data from a year-long study involving 160 students at the Cambodia Academy of Digital Technology (CADT). The dataset included attendance, interaction logs, quiz submissions, task completions, assignments, time spent on courses, and outcome scores. We utilized these data points to train and evaluate various classifiers, identifying the Random Forest classifier as the most effective. A predictive algorithm was developed using the coefficients from this classifier, tailored for practical application in educational settings. Our analysis confirmed significant correlations between the identified attributes and prediction accuracy, enhancing the algorithm's efficacy. A follow-up survey with the same participants one year later provided further validation, affirming the predictive indicators' effectiveness in improving academic performance. This comprehensive approach demonstrates the robustness of our findings and underscores the potential of predictive analytics in enhancing educational outcomes.