Personalized nutrition is emerging as an important tool in the management of chronic diseases such as obesity and type 2 diabetes. This study presents an innovative AI-driven dietary recommendation system that uses a Retrieval Augmented Generation (RAG) model to create customized smoothie recipes for people with these conditions. The system integrates the Dutch dietary guidelines of the National Institute for Public Health and the Environment (RIVM), which focus on macronutrient balance, glycemic control and sustainability. By combining the generative power of large language models (LLaMA3) with real-time queries from a curated database, the system delivers highly personalized, evidence-based recommendations. Each smoothie recipe goes through a robust validation process against nutrition and sustainability metrics to ensure it meets health guidelines for calorie, carbohydrate, fiber and fat content. The system also promotes environmental sustainability by prioritizing the use of seasonal, locally sourced ingredients. This dual focus on health and environmental sustainability makes the system a valuable tool for patients and healthcare professionals looking to optimize their diet for chronic disease management. To facilitate user interaction, a prototype web application was developed that allows the user to enter health data such as age, BMI and dietary preferences to create personalized prescriptions in real time. Future iterations of the system aim to integrate continuous health monitoring data from wearable devices to improve the accuracy and customizability of dietary recommendations.