To reach the renewable energy targets set by the European Commission, a tenfold expansion of the installed offshore wind farms is needed. Since the construction of offshore wind farms may affect local soft-sediment fauna, an efficient monitoring technique is needed to monitor the potential effects on the marine ecosystem. Here, we assess whether eDNA metabarcoding is a suitable alternative to monitor fish and epibenthos biodiversity in these difficult to access marine habitats. Water sampling and trawl surveys were conducted in parallel in 12 coastal and 18 offshore sites, the latter located inside and outside two offshore wind farms in the Belgian part of the North Sea. 12S eDNA metabarcoding retrieved 85.7% of the fish species caught in the beam trawls, whereas the COI eDNA metabarcoding only identified 31.4% of the epibenthic invertebrate species. Furthermore, the 12S marker resulted in an additional detection of 26 unique fish species, whereas the COI marker detected an additional 90 invertebrate species. Spatial patterns in alpha diversity recovered with eDNA metabarcoding were not significantly different from those observed with morphological determination. Significant differences were found in fish and invertebrate community structures between the coastal, transition and offshore zones as well as on the smaller wind farm scales, which agreed with the morphological beam trawl data. Indicator species found with morphological beam trawl monitoring for each of the three zones (coastal, transition, offshore) were also detected with 12S eDNA metabarcoding, and the latter method detected an additional 31 indicator species. Our findings show the need for adequate quality control of the obtained species lists and reveal that 12S eDNA metabarcoding analyses offers a useful survey tool for the monitoring of fish communities in offshore wind farms, but the used COI assay did not adequately capture the epibenthic communities as observed with beam trawl data.