Previous studies have observed an inverse association between alcohol consumption and multiple sclerosis (MS) risk. We aimed to investigate possible interactions between alcohol consumption, MS-associated human leukocyte antigen (HLA) genes and smoking regarding MS risk. We used a Swedish population-based case-control study (2059 incident cases, 2887 controls) matched by age, sex, and residential area. Subjects with different genotypes and alcohol consumption habits were compared regarding MS risk, by calculating odds ratios with 95% confidence intervals using logistic regression models. Interaction on the additive scale between non-drinking and both genotype and smoking were assessed by calculating the attributable proportion due to interaction (AP). There was a dose-dependent inverse association between alcohol consumption and MS risk (p for trend < 0.0001). A potentiating effect was observed between non-drinking and presence of DRB1*15:01 (AP 0.3, 95% CI 0.2-0.5) which was of similar magnitude irrespective of smoking habits. Non-drinking also interacted with smoking to increase MS risk (AP 0.2, 95% CI 0.06-0.4). Non-drinking interacts with DRB1*15:01 and smoking to increase the risk of MS. Better understanding of the mechanisms behind our findings may help to define ways to achieve protection against MS by other means than alcohol consumption.