The ornamental fish industry is a booming and emerging industry. Fish identification for fish farmers, trainers, sellers, and even buyers is an essential skill. With the rise of deep learning, object recognition is widely used in different fields. In this paper, we propose a highly flexible application via microservice architecture. It utilizes cloud computing to provide high-efficiency and low-cost identification services with NVIDIA T4 graphics processor units (GPUs) and introduces the You Only Look Once (YOLOv3) model to recognize the species of ornamental fish. Finally, mobile devices capture the photos and communicate with cloud through a representational state transfer (RESTful) application programming interface (API) to retrieve the predicted results. The proposed application identified 11 types of ornamental fish and completed each prediction within 1 s.