Background & Aims: The discovery of effective and reliable biomarkers to detect hepatitis B virus (HBV)-positive hepatocellular carcinoma (HCC) at an early stage may improve the survival of HCC. The aim of this study was to establish serum microRNA (miRNA) profiles as diagnostic biomarkers for HBV-positive HCC. Methods: We used deep sequencing to screen serum miRNAs in a discovery cohort (n=100). Quantitative polymerase chain reaction (qPCR) assays were then applied to evaluate the expression of selected miRNAs. A diagnostic 2-miRNA panel was established by a logistic regression model using a training cohort (n=182). The predicted probability of being detected as HCC was used to construct the receiver operating characteristic (ROC) curve. Area under the ROC curve (AUC) was used to assess the diagnostic performance of the selected miRNA panel. Results: The predicted probability of being detected as HCC by the 2-miRNA panel was calculated by: logit P=-2.988 + 1.299 x miR-27b-3p + 1.245 x miR-192-5p. These results were further confirmed in a validation cohort (n=246). The miRNA panel provided a high diagnostic accuracy of HCC (AUC=0.842, P<.0001 for training set; AUC=0.836, P<.0001 for validation set respectively). In addition, the miRNA panel showed better prediction of HCC diagnosis than did alpha-foetoprotein (AFP). The miRNA panel also differentiated HCC from healthy (AUC=0.823, P<.0001), and cirrhosis patients (AUC=0.859, P<.0001) respectively. Conclusions: Differentially expressed serum miRNAs may have considerable clinical value in HCC diagnosis, and be particularly helpful for AFP-negative HCC.