This paper introduces the Hybrid Adaptive Learning System, a framework that combines predictive Machine Learning with interpretable Fuzzy Logic to support personalized learning. The Machine Learning component models student performance trends, while the Fuzzy Logic module translates these predictions into transparent instructional actions. An optimization layer employing centroid defuzzification and threshold mapping ensures precise and adaptive interventions. Experiments with 120 students show that Hybrid Adaptive Learning System improves prediction accuracy, learning gains, and engagement compared to static and rule-based systems. The results demonstrate that hybrid AI approaches can provide both high performance and pedagogical explainability in adaptive learning environments.