Objective: The authors used a meta-analytic technique to quantify the evidence of an association between maternal alcohol consumption during pregnancy and childhood acute leukemia (AL), which provided a basis for the prevention of childhood AL. Methods: Relevant literatures of maternal alcohol consumption during pregnancy were comprehensively searched and screened. Subgroup meta-analysis was conducted according to the type of leukemia. Results of research data of maternal alcohol consumption during pregnancy were tested for heterogeneity. Combined OR values and 95% CIs were statistically calculated with RevMan 4.2 software; Funnel plots were applied to conduct bias analysis for those included literatures. Results: Ten related literatures were included after data screening, 4593 cases in AL group and 6157 cases in control group respectively. According to heterogeneity test result (χ2 = 16.26, P < 0.05), the combined OR values and 95% CI were calculated with random effects model, which were 1.02 (0.92-1.14), Z = 0.41, P = 0.68 > 0.05, indicating that there was no significant difference between maternal alcohol consumption during pregnancy and the risk of childhood acute leukemia (AL). Subgroup analysis: for the association between maternal alcohol consumption during pregnancy and childhood acute lymphoblastic leukemia (ALL), the combined OR value and 95% CI were 0.92 (0.84-1.00), Z = 1.92, P = 0.05, indicating that there was significant difference between two groups; for the association between maternal alcohol consumption during pregnancy and childhood acute non-lymphoblastic leukemia (ANLL), the combined OR values and 95% CI were 0.82 (0.61-1.11), Z = 1.30, P = 0.19 > 0.05, indicating that there was no significant difference between two groups. Conclusion: Maternal alcohol consumption during pregnancy is a risk factor in childhood ALL, but not in childhood ANLL. © 2010 Springer-Verlag Berlin Heidelberg.