Mass spectrometry is being widely applied to identify and quantify proteins in complex mixtures. Quantification of small molecules by integration of LC-MS extracted ion chromatogram (XIC) peaks has a long history in analytical chemistry. Similar quantification techniques applied to proteolytic protein digests have also been previously described. A comprehensive approach for label-free quantification using yeast proteome as a model have been developed. Based on spectra counts of peptides, the relative protein quantification from LC-MS/MS experiments of proteolytic protein digests was performed. Unlabeled protein samples were digested with trypsin and separated by one-dimensional nano-flow HPLC (RPLC), and mass spectra were obtained by using the survey mode of LTQ mass spectrometer with dynamic exclusion. The correlating relationship between concentrate of protein and spectra counts was confirmed. Two algorithms to normalize spectra counts ratio from different samples were compared, and the results suggested that algorithm based on average spectra count (ASC) was eximious. The method was used to biomarkers discovery in HepG2 and HepG2-HBx cell lines. The identified proteins were analyzed and classified by cluster software (Version 3.0). Finally, 107 overlap proteins were identified, among them, 9 proteins were identified up-regulated (Ratio > 1.75) and 6 proteins down-regulated (Ratio < 0.5). Further research indicated that these proteins were related with liver cancer. Altogether, the results indicated that the strategy was operable and convenient with high sensitivity and wild dynamic range, and will be significant to deliver biomarker discovery either in theory or in clinic.