In this study, we attempted to use machine learning to determine when to replace a stoma appliance. Stoma appliances must be replaced periodically, but deciding when to replace the front plate is tough. Although the front plate adheres closely to the stoma, if urine or stool leaks during use or if urine or stool gets caught between the stoma and the front plate, the front plate can peel off and cause leakage. Therefore, it is essential to perform proper replacement. However, nurses and patients who are not accustomed to exchanging them often misjudge the timing of the exchange. In this study, we attempted to use machine learning to determine when to replace a stoma appliance. By using machine learning, it is expected that the system will be able to accurately determine when to replace a stoma appliance, thereby preventing misjudgment of the timing of replacement and leakage of urine and stool. In this system, elderly stoma holders upload photos taken at home with their smartphones via the Internet, and a machine-learned database of a vast amount of past data returns whether or not it is time to replace the stoma appliance. Many stoma holders are elderly, and older people are often unfamiliar with digital devices. Therefore, by utilizing the vast database, the system aims to determine when to replace a stoma appliance with a high probability (accuracy set at 80–90%), even with out-of-focus photos (photos taken as they are).