ABSTRACT The ability of discriminating carcass characteristics of different fat cover scores of heavy carcasses, according to the European (EUS) and Brazilian (BRS) classification systems, was assessed. Fifty-six lambs, weighing between 26.25 and 46.15 kg, of the Santa Inês and crossbred Santa Inês × Dorper genetic groups were evaluated. The level of adiposity was assessed through color photography of the carcasses after refrigeration according to the EUS and BRS. The carcasses were assigned to four groups by cluster analysis according to 25 variables, namely, cold carcass weight; muscle; bone; fat; and muscle:fat and fat:bone ratios of the carcass and cuts (hindquarter, shoulder, back); kidney, pelvic, and inguinal fat; and subcutaneous fat thickness. Of the four groups obtained by cluster analysis, the scores according to the different classification systems only statistically differed between groups 1 and 4. The BRS had a higher number of variables well correlated with the scores by assessors than the groups classified by the EUS. The BRS was better correlated with tissue composition. However, most variables were better correlated with backfat thickness than the score obtained through the classification systems. Better results were obtained regarding the prediction of carcass fat by the BRS using backfat thickness or cold carcass weight. The Brazilian lamb carcass classification system better predicts tissue composition and is the best method to discriminate intermediate-fat classes when associated with cold carcass weight.