Objective: To evaluate the effects of dual-task training on static and dynamic balance in patients with multiple sclerosis. Data sources: PubMed/MEDLINE, EMBASE, Scopus, and PEDro databases were searched from inception to March 1, 2021. Methods: This study was conducted in agreement with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Two reviewers assessed studies for inclusion and extracted data. We used the Physiotherapy Evidence Database scale to evaluate the methodological quality and the risk-of-bias. Randomized clinical trial data were pooled for the meta-analysis. The effect sizes and 95% confidence interval (CI) were calculated by random-effect models. Egger regression and Begg-Mazumdar rank correlation test were used for publication bias. Results: A total of 13 studies involving 584 patients (42.3 ± 9 years mean ± SD; 377 females) met the inclusion criteria for the systematic review, while nine were included in the meta-analysis. People who received dual-task training interventions showed significant improvements in the Timed Up & Go test 0.44 [(95% CI = 0.22; 0.65), P-value<0.001], and in the Berg Balance scale 0.46 [(95% CI = 0.07; 0.85), P-value = 0.02]. Low and moderate heterogeneity between the studies was found for the Timed Up & Go test and the Berg Balance scale, respectively. Conclusion: The findings from the current meta-analysis support dual-task training as a beneficial therapy for improving dynamic balance and functional mobility in patients with multiple sclerosis. The limited number of studies that investigated static balance performance after dual-task training do not currently allow us to draw a conclusion about any possible improvements in this ability.