AbstractUnderstanding the spatiotemporal dynamics of fish species is a central concern in fish ecology and crucial for guiding management and conservation efforts. We constructed a joint species distribution model (JSDM) to simultaneously estimate the spatiotemporal distributions and densities for 21 reef fish species in the southeastern United States (SEUS). The model separately estimates encounter probability and positive density, and accounts for unobserved spatial and spatiotemporal variation using latent factors, where the correlations among species are induced. We applied the model to video data collected from a large‐scale, fishery independent survey. A clustering method was applied to the results of the JSDM to group species based on spatial and spatiotemporal synchrony in encounter probability and positive density. We found strong spatial associations among most of the reef fish species. However, species did exhibit differences in occupied habitat that varied with latitude and/or depth. Within their area of occupied habitat, almost all the species share similar spatial pattern of average density. However, for some species, annual distributions were less correlated with their expected average distributions perhaps due to differing responses to underlying spatiotemporal drivers. Some species show significant declines in abundance, for example, black sea bass, red porgy, and blueline tilefish, while a small number of species showed evidence of shifts in distribution, for example, black sea bass. The findings suggest that spatiotemporal management strategies may be of limited utility for reducing bycatch in these highly mixed reef fisheries due to high spatial correlations in occupied habitat and spatial patterns in density. Species‐specific responses to environmental change may also influence the spatiotemporal structure of reef assemblages. This work suggests management attention is needed for some of the lesser known species as they are showing declining trends in abundance.