Chronic diseases are a persistent and long lasting human health conditions that lasts for more than three months. Because of its long lasting effect chronic diseases needs a close and permanent follow-up and treatment. Today the prevalence of chronic non-communicable diseases in Ethiopia increases rapidly because different reasons like poor nutrition habit, lack physical activities, drinking alcohols, smoking and life style issues. To overcome this problem different technological applications are developed globally to support both the health professional in diagnosis process and the patients for their self-treatment activities. However, most of the apps and technologies have their own limitations like lack of sharing of information between agents, unable to reuse knowledge in the system, lack of common understandings between the computer and human beings and others. To alleviate these problems it is preferable to develop ontology-based systems that enable the above features. Ontology helps to create common understanding between human and computers, enable reusability of information, allows sharing of concepts. In this research, we have developed ontology based personalized recommendation model for diabetes patient in Ethiopian context. Context is important concept in the development of recommendation system. We have used design science research methodology in our proposed study. In the development of the proposed model, first we have developed the patient and domain or disease ontology and then the two ontologies needs to integrate in order to develop the required recommendation model. We have used Protégé ontology development tool for the development of the proposed domain or disease and patient ontology. Finally, we have evaluated the proposed model by incorporating both patients and health professional to test and analyze the performance of the model. We have evaluated both the model and the ontology. The overall performance result of the model and the ontology is 91% and 89% respectively. The evaluation result shows that the model has a good performance in giving recommendation for chronic diabetic patients