The causative agents of most coral diseases today remain unknown, complicating disease response and restoration efforts. Pathogen identifications can be hampered by complex microbial communities naturally associated with corals and seawater, which create complicating “background noise” that can potentially obscure a pathogen’s signal. Here, we outline an approach to investigate waterborne coral diseases that use a combination of coral mesocosms, tangential flow filtration, and size fractionation to reduce the impact of this background microbial diversity, compensate for unknown infectious dose, and further narrow the suspect pool of potential pathogens. As proof of concept, we use this method to compare the bacterial communities shed into six Montastraea cavernosa coral mesocosms and demonstrate this method effectively detects differences between diseased and healthy coral colonies. We found several amplicon sequence variants (ASVs) in the diseased mesocosms that represented 100% matches with ASVs identified in prior studies of diseased coral tissue, further illustrating the effectiveness of our approach. Our described method is an effective alternative to using coral tissue or mucus to investigate waterborne coral diseases of unknown etiology and can help more quickly narrow the pool of possible pathogens to better aid in disease response efforts. Additionally, this versatile method can be easily adapted to characterize either the entire microbial community associated with a coral or target-specific microbial groups, making it a beneficial approach regardless of whether a causative agent is suspected or is completely unknown.