In order to maintain the normal life of patients with depression, the home self-care is one possible way. However, how to detect the depression prone and depressive disorder to provide prevention of depression recurrence and early treatment of depression is an important issue. Therefore, this study proposed behavior pattern monitoring system for home self-care, the system for non-hospitalized patients in the treatment of depression when the home care model. The System using smart phones built-in gravity sensor to retrieve information about the patient''s behavior, and movement recognition algorithms to identify patients with patterns of behavior, and provide query system and tracking feature use to patients or relatives. With a view to the future with integrated medical information system to assist patients self-diagnose and self-care.