Non Communicable Diseases are the diseases that are not infectious but can lead to long term complications and many health disorders. Diabetes Mellitus is one among those diseases. It has become one of the major health risks that constitutes to the increasing death rates. This papers objective is to anticipate the difficulty that might be caused because of the diabetes in patients. For clustering of the dataset Partitioning Around Mediod algorithm is utilized and classification is performed by decision trees. This paper hopes for early detection of the complications which might help in timely treatment of individuals.