Objective of this research was to identify the prevalence of self - medication in the adult population in the municipality of the South of Brazil and its associated factors. It is a cross-sectional population-based study with 1,720 adults from an urban area of a municipality in the South of Brazil conducted in 2009. The research outcome was self-medication. The prevalence of self-medication was analyzed according to gender, age, monthly income per capita, skin color, schooling, use of health services and self-reported diseases and self-medication, adjusted for a complex sample. All the statistical analyzes were performed in the Stata® 11.0 program, considering the effect of the delineation of complex samples and the individual weights. The bivariate and multiple analyzes were performed through Logistic Regression. The sample consisted mostly of individuals who reported that they had not consulted with any doctor in the past 30 days and had not received a visit from ACS in the last year. The prevalence of self-medication was associated only with the variables age, health plan and medical consultation. Individuals who reported not having consulted in the last 30 days had a 40% greater chance of self-medication than those who had a medical visit.