Background: Though total shoulder arthroplasty (TSA) has been an acknowledged treatment option for glenohumeral osteoarthritis, resurfacing hemiarthroplasty (RHA) and stemmed hemiarthroplasty (SHA) may be preferred in some circumstances by surgeons, especially for treating young or active patients. However, decision-making between the RHA and SHA is controversial. Therefore, we conducted a meta-analysis to systematically compare two surgical procedures in terms of postoperative functional outcomes, range of motion (ROM), pain relief, complication rates, risk of revision. Methods: The PubMed, Embase, Web of Science and Cochrane Library were searched from inception to January 1, 2020, for all articles that compared the clinical effectiveness and safety of RHA with SHA. All eligible studies were selected based on certain screening criteria. Two investigators independently conducted the quality assessment and extracted the data. Fixed-effect and random-effect models were used for pooled results according to the degree of heterogeneity. All statistical analyses were performed by employing Stata software 14.0. Results: A total of six comparative studies involving 2568 shoulders (1356 RHA and 1212 SHA) were included in the final analysis. Patients were followed up for at least 1 year in each study. Pooled results showed that RHA was associated with a better visual analog scale (SMD 0.61, p = 0.001) but higher revision rates (OR 1.50, p = 0.016) when compared to SHA. There were no significant differences in functional outcomes, such as Constant-Murley score (SMD 0.06, P = 0.878), American Shoulder and Elbow Surgeons score (SMD 0.05, P = 0.880), Western Ontario Osteoarthritis of the Shoulder index (SMD 0.43, p = 0.258) and quick-Disabilities of the Arm, Shoulder and Hand score (SMD 0.06, p = 0.669). In addition, no differences were observed in forward flexion (SMD 0.16, p = 0.622), external rotation (SMD-0.17, P = 0.741) and overall complication rates (OR 1.42, p = 0.198).