Larval fish surveys can provide a tangible signal of the interannual variation in larval supply (recruitment). The abundance of larval fishes can vary by orders of magnitude and accurate enumeration of samples is expensive and time-consuming. Predicting the abundance of larval fish at unsampled times and locations traditionally relies on modelling based on abiotic spatial, temporal, and oceanographic factors, without considering larval fish as part of the zooplanktonic community. Larvae are typically removed from the concomitant zooplankton present in each sample, losing any understanding of ecological relationships with the surrounding assemblage. Here we use synoptically collected data from the Integrated Marine Observing System of larval fish and copepod abundance off eastern Australia to test whether incorporation of the copepod community can improve the predictive abilities of larval fish abundance models. Using larval T. novaezelandiae, S. sagax, E. australis, & S. australasicus, we demonstrate that including key copepod taxa in statistical models improves predictive ability compared to models built solely on abiotic variables. This outcome indicates the possibility for using copepod data more widely to predict the abundance of key larval fish species, an important outcome as nationally coordinated larval fish monitoring in Australia is discontinued.