Background: Obesity could be classified into two phenotypes: metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUHO). This study investigated the ability of liver enzymes to identify obesity phenotype. Methods: We conducted a cross-sectional study in 2197 obese adults (age >40 years and BMI >= 25kg/m(2)) in a rural area of central China. Results: In this population, 75% of the participants have more than one cardiometabolic risk factor. Both GGT and ALT were strongly related to the MUHO phenotype. The association between the fourth quartile of GGT and MUHO risk was strong and independent of confounder risk factors in both genders (adjusted ORs, 1.73 (95%CI 1.03-2.92) for male and 1.82 (95%CI 1.29-2.57) for female). The association between the fourth quartile of ALT and MUHO risk was strong and independent in female, but not in male (adjusted ORs, 1.65 (95%CI 0.86-3.19) for male and 1.88 (95%CI 1.29-2.75) for female). Additionally, AST was not associated with MUHO phenotype. Conclusions: Both GGT and ALT are effective markers for identifying MUHO in this population. Furthermore, the ability of GGT may be superior to ALT in male.