Small pelagic fishes are critical components of marine food webs, yet their distribution and aggregation dynamics in time and space remain unpredictable and challenging. We integrated drone-based surveys with traditional cast-net sampling to provide a comprehensive assessment of small pelagic fish assemblages in shallow tropical ecosystems in northeastern Queensland, Australia. We developed and tested a rapid drone survey protocol for detecting coastal fishes by evaluating the effects of flight altitude and camera orientation on detection success. Drone surveys detected pelagic fish aggregations dominated by mullet species, Liza vaigiensis, and longtom, Tylosurus sp., and captured aggregation behaviour and interactions with other aquatic organisms that are rarely observed with conventional sampling. In contrast, cast-net sampling provided quantitative assessment of small pelagic fish assemblage dominated by Thryssa hamiltonii and Herklotsichthys castelnaui. By combining aggregation observations from drone surveys with quantitative information from net sampling, we demonstrate that the two approaches capture complementary dimensions of tropical coastal fish assemblages. While UAVs efficiently characterise fish aggregations, activity and habitat associations, conventional sampling provides detailed information on assemblage composition and relative abundance. Integrating these fine-scale observations offers a more complete understanding of the ecological processes shaping coastal fish communities than either method alone and provides a practical framework for improving biodiversity monitoring, fisheries assessment and evidence-based management of shallow tropical coastal ecosystems.