Using data from population-based samples of adults participating in the ICS-II USA study, and using principal components analysis, we constructed oral health belief measures corresponding to the Health Belief Model (HBM) dimensions. Tests of validity and reliability were performed. Scales measuring perceived benefit of preventive practices and seriousness of oral disease had the highest validity and reliability. We used multiple regression analysis to examine sociodemographic predictors of perceived benefits of preventive practices. Race-ethnicity and age cohort were significant predictors among Baltimore and San Antonio adults. White adults and middle-aged persons in both research locations were more likely to believe in the benefit of preventive practices. Female gender, higher educational attainment, and better self-rated health were significant indicators of more positive oral health beliefs in every research location. Results also characterize persons who place lower value on preventive practices (i.e., males, less-educated persons, and those reporting poorer self-rated health). The design of effective dental public health messages and outreach efforts requires an analysis of the individual's health orientation and the factors influencing oral health beliefs. Oral health education interventions designed to improve health beliefs should contain an evaluation component for assessing the impact of education on health practices and oral health status.