Background: Early detection of head and neck squamous cell carcinoma (HNSCC) remains challenging because tissue biopsy is invasive and unsuitable for repeated monitoring or large-scale screening. This study systematically evaluated the diagnostic accuracy of human papillomavirus (HPV) DNA detection in oral gargle and rinse samples for HNSCC. Methods: This meta-analysis followed PRISMA guidelines and included 10 studies with 1315 participants. Pooled sensitivity, specificity, summary receiver operating characteristic (SROC), and diagnostic odds ratios (DOR) were calculated using random-effects models. Comprehensive subgroup analyses were conducted to explore pre-analytical and analytical heterogeneity. Results: Pooled sensitivity and specificity were 0.68 (95% CI: 0.58–0.79) and 0.91 (95% CI: 0.88–0.95), respectively, with an AUC of 0.91. Higher pooled sensitivity was observed in subgroups using smaller collection volumes (≤10 mL), non-SPF10 PCR-based detection methods, and HPV16-specific targets, whereas specificity remained consistently high across subgroups. Fagan’s nomogram showed that the test offers greater clinical value for ruling in than ruling out the disease. Conclusion: Oral HPV DNA testing via gargle and rinse samples shows high specificity but moderate, highly variable sensitivity. Therefore, it functions best as a non-invasive, supplementary tool for triage or monitoring, rather than as a standalone early detection or rule-out method for HNSCC. © 2026 Japanese Association for Dental Science. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license. http://creativecommons.org/licenses/by-nc-nd/4.0/