Bridge damage is a case in which physical forces act on the inside and outside of the bridge and maintain it in an unstable state, such as cracks, breakage, deformation, and erosion. As concrete is a brittle material, it is easily broken by external impact and can lead to serious accidents due to loss of function, so periodic management is essential. In this paper, we propose a crack management system using the embedded device for drone mounting. The embedded device for drone mounting is connected to a distance sensor, GPS module, and a camera to photograph the bridge in real time and uses a machine learning algorithm to find cracks in the bridge. When the embedded device finds a crack with a camera, it sends JSON formatted crack detection data including GPS position, timestamp, base64 encoded image data, and additional information to the crack management server over MQTT. We also propose a crack management algorithm and implemented a crack management server using a proposed algorithm. A proposed crack management server is implemented by the python web framework flask and a tensorflow machine learning library, and saves data with a mariaDB database server. The proposed system can enable fast and efficient management by using a real-time analysis method rather than a method of analyzing filmed video.