Abstract Fish ecology in urban estuaries is poorly understood. As coastal landscapes are transformed, recognizing the impact that urban structures, such as marinas, seawalls and wharfs, have on local fish populations is becoming increasingly important. The extent to which fish are able to maintain natural ecological assemblages can be measured, to a certain extent, by how closely they mimic natural habitats. In Sydney Harbour, assemblages of fish associated with artificial structures were compared with those associated with natural rocky reefs. Sampling was carried out in five locations, each with a marina, swimming enclosure and natural reef. In each location, different habitats supported different assemblages, but differences between habitats were not consistent among locations. Subsequent sampling compared artificial and natural sites in three different areas in each of three different estuaries. Results indicated that differences in fish assemblages between artificial and natural sites were greater than differences between sites within each habitat, but there were no patterns among different positions in an estuary or from estuary to estuary. This study provides initial evidence that, although artificial habitats generally support the same species as found on natural reefs, assemblages usually differed between natural and artificial habitats. In addition, without knowing if these habitats do, in fact, sustain viable populations of fish, it would be premature to label artificial structures as effective habitat for fish.