Background: Virtual reality (VR) has emerged as a therapeutic tool facilitating motor learning for balance and gait rehabilitation. The evidence, however, has not yet resulted in standardized guidelines. The aim of this study was to systematically review the application of VR-based rehabilitation of balance and gait in 6 neurologic cohorts, describing methodologic quality, intervention programs, and reported efficacy. Methods: This study follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. VR-based treatments of Parkinson disease, multiple sclerosis, acute and chronic poststroke, traumatic brain injury, and cerebral palsy were researched in PubMed and Scopus, including earliest available records. Therapeutic validity (CONTENT scale) and risk of bias in randomized controlled trials (RCT) (Cochrane Collaboration tool) and non-RCT (Newcastle-Ottawa scale) were assessed. Results: Ninety-seven articles were included, 68 published in 2013 or later. VR improved balance and gait in all cohorts, especially when combined with conventional rehabilitation. Most studies presented poor methodologic quality, lacked a clear rationale for intervention programs, and did not utilize motor learning principles meticulously. RCTs with more robust methodologic designs were widely recommended. Conclusion: Our results suggest that VR-based rehabilitation is developing rapidly, has the potential to improve balance and gait in neurologic patients, and brings additional benefits when combined with conventional rehabilitation. This systematic review provides detailed information for developing theory-driven protocols that may assist overcoming the observed lack of argued choices for intervention programs and motor learning implementation and serves as a reference for the design and planning of personalized VR-based treatments.