The rapidly expanding Atlantic bluefin tuna fattening industry is characterized by high stock densities and a high input food biomass in the form of whole bait seafood. The environmental impact of this activity must be effectively monitored within a proper sustainable development framework, to address concerns about the potential adverse effects. However, evaluation of monitoring tools for tuna farming has received less attention than other activities. Based on the principles of key taxa (Pocklington and Wells, 1992), we tested the potential use of changes in benthic macroinvertebrate assemblages, polychaetes and amphipods for this purpose. Applying a non-parametric multivariate approach for monitoring the impact of this aquaculture activity on the benthic habitat, we checked for correlations with the physicochemical environmental variables of the sediment. A hierarchical spatial design was followed, using multiple controls. Amphipods and polychaetes showed dissimilarities between impacted and control locations, with significant differences for total assemblage structure at a taxonomic level of families. Total nitrogen (TN) and total sulfur (TS) concentrations were the variables best associated with these changes for amphipods, and d 13C and total phosphorus (TP) were the best for polychaetes. However, total free sulfides (TFS) and TP were the chemical variables that best indicated the effects on sediment. Using this approach, surrogating the whole benthic assemblage to a single taxocene, our data suggest that monitoring tuna farming impact by comparing the changes in amphipod and polychaete assemblages at family level could be an optimal procedure with an excellent cost/benefit ratio.