Eating behavior is influenced by a combination of environmental and genetic factors. Although candidate gene studies have been conducted, much remains to be understood about genetic influences. Therefore, we conducted a Genome-Wide Association Study (GWAS) aims to identify new variants that influence eating disinhibition. Moreover, we test the possible association of these variants with food liking and metabolic phenotypes.We measured disinhibition in two cohorts of Italian samples using three selected statements from the Three-Factor Eating Questionnaire. Personal and clinical data were collected, as well as liking for different foods and beverages. GWAS was carried out in 1124 individuals; then the best signals (p-value < 1×10−5) were studied for replication in 426 independents participants. To study the link of eating disinhibition and associated variants with food liking and metabolic traits, we used linear mixed models and Structural Equation Modeling (SEM).A significant association with CAV1 (caveolin 1) gene (p < 5×10−8) was identified. The top SNP (rs6961694) resulted also associated with the liking for sweet foods and alcoholic beverages (p-value < 0.05). Moreover, we observed significant eQTL associations between this SNP and CAV1 expression levels in human tissues such as adipose subcutaneous tissue, pancreas and brain hippocampus (p-value = 0.00022, 0.00015 and 0.017, respectively). Although higher values of BMI, waist, hips and triglycerides were significantly associated with increasing eating disinhibition (p-value < 0.05), no association emerged between the rs6961694 SNP and anthropometric or lipids phenotypes.In conclusion, we describe a significant association between eating traits and CAV1 gene, providing new knowledge on the link existing between genetics, eating behaviour and health status.