The increasing availability of online information has introduced challenges like information overload, particularly in the context of personalized dietary management for individuals with diabetes. This study introduces a personalized food recommendation system designed to enhance diabetes management. The system employs a hybrid filtering method, combining content-based and collaborative filtering techniques, with linear regression serving as the core machine learning algorithm to improve recommendation accuracy. Evaluation results reveal that the system achieved a high accuracy rate of 98.34%, effectively supporting diabetic patients in making well-informed dietary decisions. This advancement contributes to better glycemic control and improved adherence to dietary recommendations, underscoring the system’s potential to provide personalized nutrition advice for managing diabetes. The findings of this work offer significant contributions to health informatics, paving the way for future e-health innovations in personalized nutrition.