Diabetes is a serious, long-term condition that occurs when the body cannot produce any or enough insulin or cannot effectively use the insulin it produces. Diabetes type I (DT1) – a chronic, life-long, non-curable disease with high percentage of complications. Several studies have proved that special diet, appropriate exercises and long-term lifestyle changes can help in managing of diabetes, holding it in a compensated form and decreasing complications severity. However, it’s not easy to manage DT1. All over the world exist governmental and nongovernmental organizations, called “School of Diabetes” that educate patients in how to control their disease. The aim of this paper is to propose a patient-specific, multilevel and predictive model, using meal and exercise guidelines based on a comparative analysis of modern Schools of Diabetes (USA, Russia, China), having a problem-oriented dataset built upon the Diabetes Data Set (UCI) and personal data records from a targeted group of 5 patients with DT1. Obtained results from creating 5 patient-specific models and a 3 month investigation period show that the proposed modelling technique can decrease patients’ HbA1c and heart rate values.