Students’ performance measurements are necessary for contemporary evaluation techniques, tools, and applications to monitor students and accomplishments throughout the course. Engaging students in teaching-learning activities on a regular basis has benefits and improves learning outcomes. However, there are a number of research problems when it comes to customizing and generalizing prediction models for student achievement in a course. Three significant contributions are made in this work. In order to choose predictive models for use in blended learning systems, we first present an assessment of the current literature. Secondly, this method aids in selecting crucial indicators for prediction models. Third, to improve model prediction accuracy and have a better understanding of student performance, when selecting important predictive models, reuse academic performance factors.