Coral reef communities are globally threatened by natural and anthropogenic causes. To manage coral reefs, basic data such as coral cover, etc. with acceptable accuracy are necessary. Classified satellite images created using the GPS position data of underwater objects (coral, sand, seagrass, macro algae) are obtained mainly from published sources. The locations of these data include Okinawa (Japan), Seribu Islands (Indonesia), and Great Barrier Reef (Australia). To create coral reef cover maps, we have used two satellite images (e.g. Sentinel and Planet; 10 m and 3 m resolution, respectively). We have used the maximum likelihood classification combined with the object-based classification on ArcGIS Pro. Lastly, we have assessed the accuracy of classified images based on the two satellite images. Preliminary results of these are presented here.