Due to the predicted future demand for critical metals, abyssal plains covered with polymetallic nodules are currently being prospected for deep-seabed mining. Deep-seabed mining will lead to significant sediment disturbance over large spatial scales and for extended periods of time. The environmental impact of a small-scale sediment disturbance was studied during the ‘DISturbance and reCOLonization’ (DISCOL) experiment in the Peru Basin in 1989 when 10.8 km2 of seafloor were ploughed with a plough harrow. Here, we present a detailed description of carbon-based food-web models constructed from various datasets collected in 2015, 26 years after the experiment. Detailed observations of the benthic food web were made at three distinct sites: inside 26-year old plough tracks (IPT, subjected to direct impact from ploughing), outside the plough tracks (OPT, exposed to settling of resuspended sediment), and at reference sites (REF, no impact). The observations were used to develop highly-resolved food-web models for each site that quantified the carbon (C) fluxes between biotic (ranging from prokaryotes to various functional groups in meio-, macro-, and megafauna) and abiotic (e.g. detritus) compartments. The model outputs were used to estimate total system throughput, i.e., the sum of all C flows in the food web (the ‘ecological size’ of the system), and microbial loop functioning, i.e., the C-cycling through the prokaryotic compartment for each site. Both the estimated total system throughput and the microbial loop cycling were significantly reduced (by 16% and 35%, respectively) inside the plough tracks compared to the other two sites. Site differences in modelled faunal respiration varied among the different faunal compartments. Overall, modelled faunal respiration appeared to have recovered to, or exceeded reference values after 26-years. The model results indicate that food-web functioning, and especially the microbial loop, have not recovered from the disturbance that was inflicted on the abyssal site 26 years ago.