The taxonomic sufficiency approach has been proposed as a surrogate for the typical analysis of species-abundance data, especially in conditions involving prominent pollution gradients. Here, we evaluate the use of taxonomic sufficiency with infralittoral macrobenthic data derived from samples taken in a moderate polluted subtropical environment in southeastern Brazil, analysing five taxonomic levels and including two functional levels throughout polychaete feeding guilds and trophic groups. The data were collected seasonally at nine stations and studied for two abundance data series (0.5 and 1.0 mm sieve mesh-size). The results showed a similar ordination pattern between the two sieve mesh-size, but with the 0.5 mm sieve data a different pattern was observed during austral summer. A slight loss of information was detected using genus, family, polychaete species and their feeding guilds as taxonomic/functional units. These results together with those of the cost! benefit ratio, suggested that the family level seemed to be sufficient to detect the impact caused by moderate pollution in this shallow-water, subtropical environment. In additional, through the use of feeding guilds, similar patterns are obtained. Correlation analysis showed that chlorophyll a, total organic matter, zinc, and chromium sediment content were the variables that best explained the biological pattern observed and not always the best correlation coefficient occurring at the species level. The feeding guild approach seems to be useful and generates interpretable results similar to those obtained with the species level of the whole macroinfauna. The results showed an important cost reduction in the sample processing, suggesting that it is possible to adopt a coarser taxonomic level monitoring program even in species-rich communities.