Individualized basic instruction is an important consideration in determining outcomes of early learning, but is not necessarily provided by conventional systems in a manner that is flexible and visible. To overcome these shortcomings, a hybrid AI framework is suggested where decision trees will be combined with fuzzy logic to customize the instruction to an individual, not only depending on cognitive indicators but also on behavioral indicators. The approach entails entropy-based decision tree models and fuzzy inference modeling of subjective variables such as motivation and engagement. The system was trained and tested on learner data from a large-scale educational dataset. The model reached 94.2% instructional match grade, 91.8% adaptivity of engagement, and 96.3% explainability index. In struggling learners, there was a 21.6% increase in the accuracy of the quiz and a 43.8% decrease in content switching due to their personalized reinforcement. Personalization, transparency, and curriculum alignment were found to be better than other existing systems through the proposed design. These findings reaffirm what hybrid AI systems can mean in providing effective, real-time, and explainable learning in basic education.