The roles of sphingosine in various cancers have not been fully investigated. Our aim was to identify the relationship between serum sphingosine and the risk of hepatocellular carcinoma (HCC). Serum sphingosine in 34 normal people and 73 HCC patients were reviewed retrospectively. Receiver operating characteristic curve analysis was performed to determine the cut-off values of sphingosine in the serum. Chi-square test, t test and regression analysis were used to test the association between serum sphingosine and individual clinicopathologic parameters. Serum sphingosine was higher in HCC patients (155.91±331.5 ng/mL) with normal persons as the control (30.92±29.4 ng/mL). The sphingosine threshold according to ROC curve was set at 22.5 ng/mL with a sensitivity of 74%, and a specificity of 55.9%. Meanwhile, sphingosine in HCC patients with abnormal albumin was significantly higher than that in patients with normal albumin (t=2.452, P=0.019). When HCC patients were divided into two groups serum sphingosine was negatively associated with albumin in HCC patients (χ2=4.469, P=0.035). Moreover, the logistic regression analysis showed that large tumor size (P=0.018, OR=0.13) and a low albumin (P=0.005, OR=8.856) were two independent risk factors for serum sphingosine upregulation. High AFP coupled with high serum sphingosine, high sphingosine and high AFP respectively were found in 91.2%, 75.4%, 73% of the HCC patients. These results suggest that serum sphingosine could be treated as a marker for the risk of HCC. AFP and sphingosine in the serum could be used together for HCC diagnosis.