Diabetes Mellitus (DM) is a chronic disease that occurs when the pancreas is unable to produce insulin. Expert system is a system that seeks to adopt human knowledge to computer, so computer can solve problems as is usually done by specialist or experts. Construction of expert system in this research using Expert System Development Life Cycle (ESDLC) methodology, Android operating system, Java programming languages, XML and database program using SQLite. The process of disease detection using certainty factor method. By the certainty factor method, an accurate results obtained from calculations based on weight of symptoms of the selected user can be provided. From test results, it can be concluded that the results of the application using black box method and direct testing by the expert, both the fault functions as well as the disease misdetection processes in application are not found. This application can perform early disease detection in accordance with expert recommendation symptom data. The certainty factor method was successfully implemented in this DM expert system and can provide result a correctness percentage of detection with the highest value of 90%. The response on questionnaire from users determines the validity of disease detection results.