Extensive artificial waterways have replaced natural wetlands and created new estuarine habitats on the southern Queensland coast, Australia. Economically important fish species found in adjacent natural wetlands of mangrove, saltmarsh and seagrass also occur in the artificial waterways. Stable isotope analyses (δ 1 3 C, δ 1 5 N) were used to test whether the relative importance of basal sources of energy varied for foodwebs found in artificial (canals and tidal lakes) and natural waterways. None of the fish species differed in their isotope values between artificial waterways. In contrast, isotopic signatures of snub-nosed garfish (Arrhamphus sclerolepis; Hemiramphidae) varied greatly between natural and artificial waterways, having highly enriched δ 1 3 C values (-10.5%%) in natural wetlands, demonstrating reliance on seagrass (-11.4%%), and significantly less enriched values (-19.0%%) in artificial waterways, consistent with either local algal sources (-19.8 to -20.4%%) or a mixture of seagrass and other less enriched autotrophs from adjacent natural wetlands. Isotopic signatures of sand whiting (Sillago ciliata; Sillaginidae) were also significantly more enriched in natural (-18.2%%) than artificial (-21.0%%) habitats, but means were not far enough apart to distinguish between different sources of nutrition. δ 1 3 C values of yellowfin bream (Acanthopagrus australis; Sparidae) did not differ between artificial and natural habitats (about -20%% in both). δ 1 5 N values of fish varied among habitats only for A. sclerolepis, which in artificial waterways had values enriched by 2%% over those in natural waterways. This was consistent with a shift from seagrass (relatively depleted δ 1 5 N) as a source in natural habitat to algal sources (relatively enriched δ 1 5 N) in artificial habitats. This study provides some of the first evidence that at least some fish species rely on different autotrophs in artificial waterways than in adjacent natural wetlands.