BACKGROUND: Prior research on childhood predictors of cigarette smoking tends to focus on the prevalence rather than the quantity of smoking and rarely examined these predictors separately for smokers. Also scarce is cross-national research synthesizing the effects of childhood predictors. METHODS: Using survey data from the Global Flourishing Study of 202,898 adults, weighted to be nationally representative of populations in 21 countries and one territory, we created continuous and binary measures of daily cigarette consumption in adulthood. The binary measure of daily smoking was regressed on childhood and demographic variables for the total sample, and the continuous measure was analyzed for the sample of smokers. RESULTS: Random effects meta-analysis provides evidence that childhood maternal and paternal relationship quality (both total and smoker samples), religious service attendance (smoker sample), and being foreign-born (both samples) predict a lower likelihood of adult smoking, whereas being raised by a divorced parent (total sample), having been abused and/or an outsider in the family (both samples), and poor health growing up (both samples) predict a higher likelihood. Although effects are generally weak and mixed in some cases, their direction and strength tend to be consistent between the two samples as well as alternative measures of smoking with some exceptions. Overall, our findings are moderately robust against potential unmeasured confounding, while the effect sizes vary across countries. CONCLUSIONS: The present study offers an important new set of global findings based on a large-scale cross-national study of daily cigarette smoking and country-specific variations.