Diabetes is a metabolic disease that is characterized by high glucose level in the blood. It is a major problem affecting millions of people nowadays. Regular testing and accurate determination of glucose levels is essential for diagnosis and treatment of diabetes. Though the test involving the collection of blood sample from finger pricking may not pose any risk to a healthy adult, but it can be very painful to the diabetic patients. A noninvasive, accurate, easy-to-use and low cost diagnostic tool for diabetes is on high demand. Acetone in the exhaled breath can be estimated for the detection of blood glucose levels. Artificial Neural Network can be used to calculate the glucose levels. It is a non-invasive technique that measure blood glucose levels and the discomfort to the patients can be minimized. Finally, to provide a global connectivity Internet of things can be enabled.