Aquatic ecosystems are under increasing anthropogenic pressure, even World Heritage National Parks such as Australia’s Kakadu are not immune. To better characterise freshwater fish communities, while minimising the biases and ethical issues associated with traditional methods, there has been a shift towards environmental (e)DNA metabarcoding. In tropical lentic systems, seasonal weather drives water quality changes, in particular, the resuspension of particulate matter and primary productivity. These attributes constrain the volume of water that can be sampled, potentially affecting the detection capacity of eDNA metabarcoding. Hence, information about the volume of water samples required for robust fish detection is critical prior to developing eDNA biomonitoring programs. To examine the influence of volume on fish detections we sampled 1, 2, 5, and 10 L water samples from two distinct systems. Utilising two mt-rRNA primers (12S and 16S) and MinION sequencing, 28 fish taxa were detected. Despite both systems containing unique fish communities (PERMANOVA: F = 5.73, P < 0.001), neither taxa richness nor composition were influenced by sample volume. Several species rarely observed using traditional monitoring approaches were also detected, including Porochilus rendahli and Oxyeleotris lineolata. In the studied systems, our findings indicate that sampling at lower volumes, thereby minimising the potential issue of filter clogging, is suitable for fish detection. Our study provides the foundations for a routine eDNA fish monitoring method that is optimised for Australian tropical lentic freshwaters, whose adoption has the capacity to aid in the ongoing management of fish communities.