Remote underwater video (RUV) cameras are commonly used in marine environments to sample fish, but to date have infrequently been used in freshwater environments. We tested the utility of a range of underwater video camera deployments (floating/benthic and fixed/moving) for surveying fish assemblages within tropical wetland habitats of Kakadu National Park, Australia. RUV cameras in both fixed and moving configurations were able to detect a range of fish species in various habitat types, with shy and/or flighty species readily detected. Cameras also performed well in a variety of conditions, although dense macrophytes posed a challenge for all video deployment types. RUV camera survey data compared favourably to two traditional fish survey techniques (visual census and pop netting). However, fixed camera deployments more closely resembled data collected via traditional techniques; whereas moving RUV deployments were consistently lower than any other method. We find that RUV cameras are a valuable monitoring and sampling technique for assessing freshwater fish assemblages in tropical wetlands, and are especially useful where remoteness, difficult access, variable water quality, complex habitats and interactions between humans and dangerous aquatic life can limit the use of other traditional fish sampling techniques.