The arbitration exercise in a soccer game requires high physical fitness and all federations apply physical tests to referees, including anthropometric tests, classifying them as fit or not for the role. The aim of this study was to assess the validity of the total body fat percentage (%BF) through different evaluation methods of body composition referenced in a four-compartment (4C) model. Cross-sectional study performed in 2018 with 21 elite male referees. %BF was estimated by 4 methods: anthropometry; bioelectrical impedance analysis (BIA); Dual X-ray absorptiometry (DXA) and air displacement plethysmography (ADP). Moreover, three and four-compartment (3 and 4C) models were calculated. Bland–Altman and intraclass correlations (ICC) analysis were performed to determine validity of all methods compared to a 4C reference. The results of one-way ANOVA revealed that there was no significant difference (F=1.541; p=0.182) between %BF analyzed by 4C model (15.98 ± 6.20), anthropometry (mean ± SD, 18.46 ± 7.03), ADP (16.19 ± 6.24), BIA (16.67 ± 5.30), DXA (20.33 ± 6.56) and 3C (16.92 ± 5.53). The Bland–Altman analysis showed that all methods analyzed overestimate %BF compared to the 4C model. The best agreement was obtained from the ADP evaluation (bias=-0.2), followed by BIA (bias=-0.6), 3C (bias=-0.9), anthropometry (bias=-2.4) and DXA (bias=-4.3). Validation assessed by ICC was excellent (ICC≥0.90) in most methods, except for anthropometry (ICC=0.80) and DXA (ICC=0.71). Overall, the results suggest that ADP, BIA and 3C were the best method to %BF evaluation. Nevertheless, anthropometry remains as a feasible method to monitor %BF of elite soccer referees.