In this paper, we designed a patient-centric cloud-based diabetes lifestyle management system. It is composed of three layers, namely sensing, communication and user interface. The goal of this cloud-based diabetes lifestyle management system is to provide type-2 diabetes mellitus patients useful information to remind user's blood sugar level. The function of the sensor networks in this framework is to collect the data from human body and human activity as well as environmental information that may have effects on the healthy statement of the diabetes patients. The communication and central server part will handle the data exchange and data analysis that help to generate a final decision data and sent to user interface to remind the user of valuable information. Different from traditional m-health system, the presented approach provides a rule algorithm which enables the rescue decision in the cloud server and transmits through the communication level, and finally provides an integrated user interface for diabetes users. An early warning user interface for diabetes patients has been designed and presented in this paper. The evaluation results show that diabetes has higher occurrence of people in elder groups who have smoking habitant, and less social activities.