Abstract: Classification models of hepatocellular carcinoma (HCC) patients and healthy people were built from mass spectrometry data, which could be used for the detection of HCC. Surface-enhanced laser desorption and ionization time-of-flight mass spectrometry (SELDI-TOF-MS) technique was applied to get the data of serum protein from HCC patients and healthy people. Then PLS variable selection method was used to deal with the data and to establish the classification model. The cross validation relativity coefficients of the model cames to over 0.96. Furthermore the important factors or variables that discriminated HCC patients and healthy people were found by analyzing the model. The 30 variables were several peak intensities of protein from some m/z sections, which could express the up-regulation or down-regulation of protein in the sections. As potential biomarkers, the proteins may be closely related to the formation of HCC, which can be deeply studied. The classification figures constructed by the fitting value of the model in the article are clear and intuitive, and can express the discrimination effect of the model well.