Objectives: To assess the diabetic patient condition using different diabetic scales and its relation to the blood glucose level and to assess the impact of health education message on their knowledge and to evaluate the possibility to create computational models for assessing diabetic patient condition automatically. Method: An interventional study was carried out on diabetic patients attending the diabetic clinic of Mansoura General Hospital, during the period of May to September 2015. Results: Random blood sugar showed significant positive correlation with Diabetes duration, BMI, depression and significant negative correlation with self-efficacy, empowerment, and social support, stage of change for diet and stage of change for exercise. However, diabetes activity showed significant positive correlation with different scales (Diabetes duration, BMI, social support, self-efficacy, empowerment, stage of change for diet and stage of change for exercise) except with random blood glucose and depression which showed significant negative correlation. Linear regression analysis showed that stage of change for exercise, social support, empowerment and BMI are common predictors for both random blood glucose and diabetes activity. Based on this study, two computational models could be concluded for predicting diabetes activity and random glucose level from different diabetes scales using artificial neural networks (ANNs). Improvement in the knowledge was observed after the health education setting as detected by the increase in the mean knowledge score. Conclusion: Computational models for diabetes condition assessment could be concluded where different variables under study were found related to the outcome of diabetes either management adherence or blood glucose level.