Background: Stroke-related balance and gait impairments significantly reduce independence, mobility, and quality of life. Virtual reality (VR)–based rehabilitation has emerged as a promising strategy to enhance motor recovery by integrating multisensory feedback, task-specific training, and high patient engagement. Objective: To synthesize randomized controlled trial (RCT) evidence on the effects of VR-based rehabilitation on balance, gait, and functional outcomes in post-stroke populations. Methods: A systematic search of PubMed, Scopus, Web of Science, CINAHL, and CENTRAL (2016–2025) identified RCTs involving VR interventions—immersive, semi-immersive, or non-immersive—delivered alone or combined with conventional physiotherapy. Eligible studies assessed at least one quantitative balance, gait, or functional mobility outcome. Two reviewers independently screened, extracted data, and assessed methodological quality using the PEDro scale. Results: Nine RCTs (316 participants) met inclusion criteria. VR modalities included Nintendo Wii–based programs, immersive head-mounted displays, VR-augmented robotic-assisted gait training, and treadmill-based VR gait training. Eight of nine trials demonstrated significant balance gains (e.g., Berg Balance Scale, Functional Reach Test) in VR groups compared with controls. Most studies also reported gait improvements (Timed Up and Go, 10-meter walk velocity, Dynamic Gait Index) and functional benefits in daily living activities, trunk control, and lower limb motor function. VR was generally well tolerated with high adherence and no serious adverse events. One large trial found VR outcomes comparable to high-intensity non-VR treadmill training. Conclusion: VR-based rehabilitation is effective for improving balance, gait, and functional outcomes after stroke, particularly when integrated with conventional therapy. Variability in VR systems, dosage, and outcome measures warrants further research to establish standardized protocols.