The epidemic of diabetes in American Indian (AI) communities is a serious public health challenge. The incidence and prevalence of diabetes have increased dramatically with accompanying increases in body weight and diminished physical activity. Daily diabetes care is primarily handled by the patients and their families, and the effectiveness of diabetes control is largely impacted by self-care strategies and behaviors. Thanks to the quasi-ubiquitous use of cell phones in most AI tribes, in this paper we propose a cell phone- based proactive diabetes self-care system, MobiDiaBTs. It is customized for AI patients using a personalized approach that considers the unique social, cultural, political, and demographic characteristic of AIs. The platform effectively and automatically collects users' physical and social behavior data and offers real-time diabetes health recommendations. It also can help a patient to interact with fellow patients in a trust-worthy and privacy-preserving environment.