An underwater camera assembly is configured to be dragged behind a boat underwater. The camera assembly includes a housing having a hollow interior that is configured to hold a camera in waterproof manner. The camera assembly includes at least one rail disposed along an outer surface of the housing and being configured to slidingly receive and interlockingly engage a first accessory (camera positioning fin). The camera assembly can further include a second accessory (trolling fin) that is also interlockingly, yet releasably, engaged to the housing.