OBJECTIVE: This study aimed to elucidate the efficacy, safety, and long-term implications of vagus nerve stimulation (VNS) as a viable therapeutic option for patients with upper limb dysfunction following a stroke. METHODS: Data from the following libraries were searched from inception to December 2022: PubMed, Wanfang, Scopus, China Science and Technology Journal Database, Embase, Web of Science, China Biology Medicine Disc, Cochrane Library, and China National Knowledge Infrastructure. Outcomes included indicators of upper limb motor function, indicators of prognosis, and indicators of safety (incidence of adverse events [AEs] and serious AEs [SAEs]). Two of the authors extracted the data independently. A third researcher arbitrated when disputes occurred. The quality of each eligible study was evaluated using the Cochrane Risk of Bias tool. Meta-analysis and bias analysis were performed using Stata (version 16.0) and RevMan (version 5.3). RESULTS: Ten trials (VNS combined with rehabilitation group vs. no or sham VNS combined with rehabilitation group) with 335 patients were included in the meta-analysis. Regarding upper extremity motor function, based on Fugl-Meyer assessment scores, VNS combined with other treatment options had immediate (mean difference [MD] = 2.82, 95% confidence interval [CI] = 1.78-3.91, = 62%, < 0.00001) and long-term (day-30 MD = 4.20, 95% CI = 2.90-5.50, < 0.00001; day-90 MD = 3.27, 95% CI = 1.67-4.87, < 0.00001) beneficial effects compared with that of the control treatment. Subgroup analyses showed that transcutaneous VNS (MD = 2.87, 95% CI = 1.78-3.91, = 62%, < 0.00001) may be superior to invasive VNS (MD = 3.56, 95% CI = 1.99-5.13, = 77%, < 0.0001) and that VNS combined with integrated treatment (MD = 2.87, 95% CI = 1.78-3.91, = 62%, < 0.00001) is superior to VNS combined with upper extremity training alone (MD = 2.24, 95% CI = 0.55-3.93, = 48%, = 0.009). Moreover, lower frequency VNS (20 Hz) (MD = 3.39, 95% CI = 2.06-4.73, = 65%, < 0.00001) may be superior to higher frequency VNS (25 Hz or 30 Hz) (MD = 2.29, 95% CI = 0.27-4.32, = 58%, = 0,03). Regarding prognosis, the VNS group outperformed the control group in the activities of daily living (standardized MD = 1.50, 95% CI = 1.10-1.90, = 0%, < 0.00001) and depression reduction. In contrast, quality of life did not improve ( = 0.51). Safety was not significantly different between the experimental and control groups (AE = 0.25; SAE = 0.26). CONCLUSION: VNS is an effective and safe treatment for upper extremity motor dysfunction after a stroke. For the functional restoration of the upper extremities, noninvasive integrated therapy and lower-frequency VNS may be more effective. In the future, further high-quality studies with larger study populations, more comprehensive indicators, and thorough data are required to advance the clinical application of VNS. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/, identifier: CRD42023399820.