The lack of an accurate biomarker in hepatocellular carcinoma (HCC) has hindered early detection, diagnosis, and treatment. Circular RNAs (circRNAs), which can be used as novel biomarkers in liquid biopsies, have been brought to light as a result of the advances in research on molecular biomarkers and the progression of genomic medicine. We conducted a meta-analysis of the diagnostic accuracy of serum/plasma circRNAs or the combination of circRNAs and α-fetoprotein (AFP) in HCC. We identified eight studies that met the inclusion/exclusion criteria from PubMed, Web of Science, EMBASE, and Cochrane Library databases. The data were pooled, and the sensitivity, specificity, diagnostic odds ratio (DOR), positive likelihood ratio (+LR), and negative likelihood ratio (-LR) with 95% confidence intervals (CIs) were calculated. The areas under the summary receiver operator characteristic (SROC) curves (AUCs) were also calculated. The sensitivity of circRNAs was 0.82 (95% CI: 0.78–0.85), and the specificity was 0.82 (95% CI: 0.78–0.86). The sensitivity of AFP was 0.65 (95% CI: 0.61–0.68), and the specificity was 0.90 (95% CI: 0.85–0.93). The AUC was 0.89 (95% CI: 0.86–0.91) for circRNAs and 0.77 (95% CI: 0.74–0.81) for AFP. The sensitivity of the combination of circRNAs and AFP was 0.88 (95% CI: 0.84–0.92), specificity was 0.86 (95% CI: 0.80–0.91), and AUC was 0.94 (95% CI: 0.91–0.96). Additionally, a subgroup analysis was conducted based on the control groups used; the diagnostic accuracy was particularly high in the comparison of HCC vs. healthy controls. In summary, serum/plasma circRNAs are accurate biomarkers suitable for clinical use for detecting HCC, and the combination of circRNAs and AFP improved the diagnostic accuracy.