This meta-analysis aims to quantitatively synthesize all available data on the association between tobacco smoking and multiple myeloma (MM) risk. Eligible studies were identified and pooled effect estimates (odds ratios and relative risks) were calculated regarding ever, current and former smoking. Separate analyses were performed on case–control and cohort studies, as well as on males and females. Meta-regression analysis with percentage of males, mean age, years of smoking, pack-years, cigarettes per day, years since quit and age at onset was performed. Forty articles were deemed eligible; of them 27 used a case–control design (4,625 cases and 21,591 controls) and 13 used a cohort design (2,228 incident cases among a total cohort size equal to 1,852,763 subjects). Ever smoking was not associated with MM risk (pooled effect estimate = 0.92, 95% confidence interval (CI): 0.85–1.00); similar results were obtained for current (pooled effect estimate = 0.87, 95% CI: 0.74–1.03) and former smoking (pooled effect estimate = 1.04, 95% CI: 0.96–1.13). Regarding ever smoking, the null association was reproducible upon cohort studies (pooled effect estimate = 1.01, 95%CI: 0.89–1.15), whereas the inverse association in case–control studies (pooled effect estimate = 0.87, 95% CI: 0.78–0.96) was particularly due to the bias-prone hospital-based ones. Meta-regression analysis did not yield statistically significant results. In conclusion, MM does not seem to be associated with tobacco smoking. There is a need to further explore how molecular mechanisms are involved in the resistance of MM progenitor cells toward smoking.