Background/Aims: Circulating (serous or plasmic) long non-coding RNA (lncRNA) as biomarkers for predicting the diagnosis or prognosis of human disease have been well documented. Due to the sensibility or specificity limitation of Alpha Fetoprotein (AFP), a cluster lncRNAs were revealed as fingerprints for hepatocellular carcinoma (HCC). In this study, we enrolled all the reported circulating lncRNAs in HCC as candidate targets and examined in an independent cohort. Methods: The candidate lncRNAs were determined by qRT-PCR divided into training and validation sets. The risk score analysis was employed to evaluate the potential diagnosis ability of the lncRNAs independently or combining with AFP value. The receiver operating characteristic curve (ROC) was applied for presentation of sensibility or specificity. Results: Among the ten candidate circulating lncRNA, LINC00152, RP11-160H22.5, XLOC014172 and LOC149086 were screened with significant difference in training set. Further investigation in validation set indicated LINC00152, RP11-160H22.5 and XLOC014172 might be the fingerprints for HCC comparing with chronic hepatitis (CH) patients or healthy controls. The risk score analysis revealed the combination of three lncRNAs with AFP could distinguish the HCC from either CH or healthy control with the area under curve value (AUC) of 0.986 and 0.985, respectively. Conclusion: The three lncRNAs may act as novel biomarkers for acting as fingerprint in HCC combining with AFP.