OBJECTIVES: The aim of the present study was to examine the constructive attitudes towards improving eating habits and their relationship with the prevalence of metabolic syndrome (MetS) using data from the National Health and Nutrition Survey (2005) and the Comprehensive Survey of Living Conditions (2005). METHODS: Individuals aged >30 years (N=3084) were enrolled in this study. Exploratory factor analyses were performed to examine the constructive attitudes towards improving eating habits by using 14 parameters. In addition, confirmatory factor analysis was performed and the Cronbach α value was calculated. Furthermore, the relationship between attitudes towards improving eating habits and MetS prevalence was examined according to gender by using multinomial logistic regression analysis, after adjusting for age, number of members in a household, exercise habits, and rest by sleeping. RESULTS: Two factors were extracted: balanced diet (BD) and control of food intake (CFI) (Cronbach α, 0.82 and 0.75, respectively). The goodness of fit model, based on the structural equation models, was adequate (goodness of fit, 0.96). No relationship was noted between BD and MetS prevalence. The MetS prevalence differed according to gender: "Strongly suspected MetS (sure MetS)", 16.3% (male, 24.4%; female, 10.7%), "Preliminary MetS (pre MetS)", 15.2% (male, 24.1%; female, 8.9%), "non-suspected MetS (non MetS)", 68.5% (male, 51.5%; female, 80.5%). The odds ratios (95% confidence interval) for pre MetS and sure MetS were 0.57 (0.42-0.78) and 0.52 (0.38-0.71) in males and 0.36 (0.25-0.53) and 0.39 (0.27-0.56) in females, respectively, when non MetS was used as a reference. CONCLUSION: The results of this study indicated that the attitudes towards improving eating habits primarily focused on BD and CFI. Therefore, improving eating habits towards in terms of better CFI would be effective in preventing MetS in both genders.