Several studies have focused on identifying the significant behavioral predictors of learning performances in web-based courses by examining the log data variables of learning management systems, including time spent on lectures, the number of assignments submitted, and so forth. However, such studies fail to quantify the impact of emotional, motivational, behavioral, and cognitive–metacognitive factors simultaneously. This research was an attempt to understand the relations between students’ motivation, cognitive–metacognitive strategies, behavior, and learning performance in the context of blended courses in higher education. Then, relevant predictors are used to obtain a model to classify the students’ performance and to identify those who are at risk of failing the course. The authors conducted an empirical study in a higher educational course with 137 Mexican students. Nineteen variables related to emotions, motivation, cognitive–metacognitive strategies, and behavior. Only six were found to be significant. These variables explain approximately 67% of the variance between each student’s overall grade. The model, based on those variables, correctly classifies 96% of the students.