OBJECTIVE: To estimate the prevalence and predictors of illicit drug use among school-going adolescents in Harare, Zimbabwe. METHODS: We used data from the Global School-based Health Survey (GSHS) conducted in 2003 in Harare to obtain frequencies of a selected list of characteristics. We also carried out logistic regression to assess the association between illicit drug use and explanatory variables. For the purpose of this study, illicit drug use was defined as marijuana or glue use. RESULTS: A total of 1984 adolescents participated in the study. Most of the sample were females (50.7%), 15-year- olds (30.3%), nonsmokers and non-alcohol drinkers. Nine percent of the subjects (13.4% males and 4.9% females) reported having ever used marijuana or glue. Males were more likely to have used marijuana or glue than females (OR=2.70; 95% CI [1.47, 4.96]). Marijuana or glue use was positively associated with cigarette smoking (OR=11.17; 95% CI [4.29, 29.08]), alcohol drinking (OR=7.00; 95% CI [3.39, 14.47]) and sexual intercourse (OR=5.17; 95% CI [2.59, 10.29]). Parental supervision was a protective factor for marijuana or glue use (OR=0.31; 95% CI [0.16, 0.61]). CONCLUSIONS: Public health intervention aimed to prevent marijuana or glue use among adolescents should be designed with the understanding that illicit drug use may be associated with other behaviors such as teenage sexual activity, cigarette smoking and alcohol use.