The Mediterranean (MED) diet was associated with a reduced risk of chronic disease, but the epidemiological studies reported inconsistent findings related to the MED diet and non-alcoholic fatty liver disease (NAFLD) risk. This age and the gender-matched case-control study were conducted among 247 adult patients. The MED diet score was obtained based on the Trichopoulou model. Multivariate logistic regression was used to examine the association between the MED diet and NAFLD risk. NAFLD prevalence in people with low, moderate and high adherence to the MED diet was 33, 13⋅1 and 4⋅6 %, respectively. The increasing intake of the MED diet was significantly related to the increment intake of nuts and fruits, vegetables, monounsaturated fatty acid/polyunsaturated fatty acid ratio, legumes, cereals and fish. However, total energy consumption, low-fat dairy and meats intake were reduced (P for all < 0⋅05). Following control for age, the person in the highest of the MED diet tertile compared with the lowest, the odds of NAFLD decreased (OR: 0⋅40, 95 % CI: 0⋅17–0⋅95). This relation became a little stronger after further adjusting for sex, diabetes, physical activity and supplement intake (OR: 0⋅36, 95 % CI: 0⋅15–0⋅89). However, this association disappeared after adjusting for body mass index, waist and hip circumference (OR: 0⋅70, 95 % CI: 0⋅25–1⋅97). High adherence to the MED diet was associated with a 64 % reduction in NAFLD odds before some anthropometric variable adjustments. However, further prospective studies are required, particularly in BMI-stratified models.