Monitoring of artificial reefs (ARs) has been conducted through such methods as visual censuses, surveys using fishing gear, and echo sounder. These methods have disadvantages: visual census is not possible at ARs in deeper waters, fishing gear surveys are invasive to fish individuals, and echo sounders have difficulty in species identification. A new AR monitoring method is required to compensate for these disadvantages. While eDNA has become a valid monitoring tool for marine biodiversities, it is influenced by degradation and transport of the molecules that affect information about the spatio-temporal distribution of fish. An understanding of the relationship between current fields and eDNA distribution, particularly in open waters, is critical when using eDNA as an index for fish aggregation at ARs. We investigated the relationship between eDNA distribution and current fields around an AR for four dominant species (Engraulis japonicus, Parapristipoma trilineatum, Scomber spp and Trachurus japonicus) in Tateyama Bay, Japan. The highest density of fish schools is formed directly above or at the upstream side of ARs. If we assume that the center of eDNA originates at these locations at an AR and eDNA is simply transported by currents, a higher density of eDNA would distribute downstream from the AR. However, our results indicate that eDNA distribution is in accord with actual fish distribution, namely eDNA densities are more abundant in the upstream side of ARs. We thus consider that eDNA distribution is more influenced by actual distribution patterns than by the transport processes.