Bladder cancer is a common and recurrent urologic malignancy. Although cystoscopy remains the diagnostic gold standard, its invasiveness, high cost, and limited patient compliance restrict its routine application. Non-invasive biomarkers have emerged as promising alternatives; however, their diagnostic performance has not yet been systematically compared across studies.We searched PubMed, Embase, Web of Science, and the Cochrane Library for publications available until February 21, 2025. A total of 26 original studies on non-invasive diagnostic methods for bladder cancer were included. A Bayesian network meta-analysis was performed to compare biomarkers derived from urine and blood samples. Diagnostic performance was evaluated using sensitivity, specificity, diagnostic odds ratio (DOR), and the superiority index. Subgroup analyses were conducted for microRNAs and combined biomarker strategies. Study quality was assessed with the QUADAS-2 tool, and model convergence was verified using Rhat values.Significant differences in biomarker performance were identified. In urine samples, angiogenin achieved the highest superiority index (5.28), while miR-125b was the best-performing microRNA (10.97). Combined detection strategies involving TERT/FGFR3/TP53/PIK3CA/KRAS demonstrated strong performance (8.54). In blood samples, miR-181b-5p and fibronectin had an index of 3.02, whereas miR-301a-3p exhibited the greatest superiority (50.71).This study is the first to systematically compare non-invasive bladder cancer biomarkers within a Bayesian framework. Specific microRNAs and combined detection strategies demonstrated robust diagnostic potential, providing a promising alternative to cystoscopy, particularly for early screening and patient monitoring.https://www.crd.york.ac.uk/PROSPERO/, identifier CRD420251018161.Copyright © 2025 Zhang, Chen, Ye, Bai and Han.