Abstract Background Studies regarding whether light to moderate alcohol consumption is associated with a lower risk of cardiovascular diseases (CVD) have generated mixed results. Further, few studies have examined the potential impact of alcohol consumption on diverse disease outcomes simultaneously. We aimed to prospectively study the dose-response association between alcohol consumption and risk of CVD, cancer, and mortality. Methods This study included 83,732 adult Chinese participants, free of CVD and cancer at baseline. Participants were categorized into 6 groups based on self-report alcohol consumption: 0, 1–25, 26–150, 151–350, 351–750, and > 750 g alcohol/wk. Incident cases of CVD, cancers, and mortality were confirmed by medical records. Hazard ratios (HRs) for the composite risk of these three outcomes, and each individual outcome, were calculated using Cox proportional hazard model. Results During a median follow-up of 10.0 years, there were 6411 incident cases of CVD, 2947 cancers and 6646 deaths. We observed a J-shaped relation between alcohol intake and risk of CVD, cancer, and mortality, with the lowest risk at 25 g/wk., which is equivalent to ~ 2 servings/wk. Compared to consuming 1–25 g/wk., the adjusted HR for composite outcomes was 1.38 (95% confidence interval (CI):1.29–1.49) for non-drinker, 1.15 (95% CI: 1.04–1.27) for 26–150 g/wk., 1.22 (95% CI: 1.10–1.34) for 151–350 g/wk., 1.33 (95% CI: 1.21–1.46) for 351–750 g/wk., and 1.57 (95% CI: 1.30–1.90) for > 750 g/wk., after adjusting for age, sex, lifestyle, social economic status, and medication use. Conclusions Light alcohol consumption at ~ 25 g/wk was associated with lower risk of CVD, cancer, and mortality than none or higher consumption in Chinese adults.