Prawn trawling is known to have deleterious localised effects on benthic ecosystems, particularly slow growing and sessile organisms, although long-term impacts are varied and less certain for relatively motile bycatch, such as fish. This study examined changes in the community structure and biomass of fish bycatch in trawl surveys from Spencer Gulf, which has supported a prawn fishery since 1967. With the view of augmenting ecosystem-based management in the future, two fishery-independent by-catch surveys were undertaken throughout the gulf six years apart, enabling a long-term temporal comparison of trawl by-catch. Differences in the community structure and biomass of fish were found to be driven primarily by regional (i.e. latitude) and/or inter-survey differences, while trawl intensity (i.e. hours trawled per km2) had no consistent influence across the two surveys. The absence of a consistent trawl-related distribution in fish biomass between the two surveys suggests our ability to detect any localised effects of trawling is potentially limited to discrete time periods. Temporal declines in fish biomass were evident across multiple species and size classes; however, they were most prominent in the southern half of the gulf, where trawl effort is historically low. While variable recruitment and/or migration within southern Spencer Gulf might explain these temporal decreases in the biomass, it is important to maintain similar ecosystem-focused studies to increase our understanding of potential trawl-related impacts. Enhancing this understanding will require (1) in-depth biological and ecological information about potential indicator species from the bycatch and (2) where possible, greater temporal resolution (i.e. intra-annual) and replication (i.e. inter-annual) to develop accurate biological reference points.