Aims To determine (i) whether the strength of Australian alcohol control policy in three domains (youth access, trading hours and drink driving) changed during the 2000s; and (ii) estimate associations between these policies and adolescent drinking after adjusting for television alcohol advertising exposures, alcohol outlet density, alcohol price changes, exposure to negative articles about alcohol in daily newspapers and adult drinking prevalence. Design Repeated cross-sectional surveys conducted triennially from2002 to 2011. Multi-levelmodelling examined the association between alcohol control policies and drinking prevalence after adjusting for covariates. Setting Four Australian capital cities between 2002 and 2011. Participants Students aged 12-17 years participating in a triennial national representative school-based survey (sample size range/survey: 9805-13 119). Measurements Outcome measures were: past month drinking and risky drinking (5+ drinks on a day) in the past 7 days. Policy strength in each of three domains (youth access, trading hours, drink-driving) were the key predictor variables. Covariates included: past 3-month television alcohol and alcoholcontrol advertising, alcohol outlet density, alcohol price change, negatively framed newspaper alcohol articles, adult drinking prevalence and student demographic characteristics. Findings During the study period, the strength of youth access policies increased by 10%, trading hours policies by 14% and drink-driving policies by 58%. Past-month and risky drinking prevalence decreased (e.g. past-month: 2002: 47.4% to 2011: 26.3%). Multivariable analyses that included all policy variables and adjusted for year, student and other covariates showed past-month drinking to be associated inversely with stronger trading hours policies [odds ratio (OR) = 0.80, 95% confidence interval (CI) = 0.69, 0.94], but not youth access (OR = 0.92 95% CI = 0.81, 1.04) or drink-driving (OR = 1.00, 95% CI = 0.93, 1.09). Risky drinking was associated inversely with stronger youth access policies (OR = 0.82, 95% CI = 0.69, 0.98), but not trading hours (OR = 0.85, 95% CI = 0.66, 1.09) or drink-driving (OR = 1.02, 95% CI = 0.90, 1.14) policies. Conclusions Population-directed policies designed to reduce alcohol availability and promotion may reduce adolescents' alcohol use.