Aim: To investigate the association between prolonged pacifier use during childhood and smoking in adolescence and early adulthood. Methods: A historical cohort study including patients from a dental private office was designed. Dental records were used, which contained complete data about sucking habits from 314 children (2-10 years of age) who had attended a private dental office from 1988 to 1994 in Ibiá, Brazil. Then, we collected data about the smoking habits from 261 subjects who were successfully contacted again from 2004 to 2006. Our outcome variable was smoking, and subjects who had smoked more than 100 cigarettes during their lifetime were classified as smokers. Poisson regression analysis matched the association between oral habits and smoking. Incidence rate ratios (IRR) and 95% confidence intervals (95% CI) were calculated. The level of significance was set at 5%. Results: We observed a statistically significant association between prolonged pacifier use (more than 24 months) and smoking (IRR = 4.48; 95% CI 2.32-8.65). Breastfeeding, in contrast, was a protective factor (IRR = 0.64; 95% CI 0.42-0.96). Conclusions: Prolonged pacifier use during childhood is positively associated with smoking initiation in adolescence and early adulthood.