Objective To identify potential serum biomarkers that could be used to discriminate lungcancers from normal. Methods Proteomic spectra of twenty-eight serum samples from patientswith non-small cell lung cancer and twelve from normal individuals were generated by SELDI(Surfaced Enhanced Laser Desorption/Ionization) Mass Spectrometry. Anion-exchange columns wereused to fractionate the sera into 6 designated pH groups. Two different types of protein chip arrays,IMAC-Cu and WCX2, were employed. Samples were examined in PBSII Protein Chip Reader(Ciphergen Biosystem Inc) and the discriminatory profiling between cancer and normal samples wasanalyzed with Biomarker Pattern software. Results Five distinct potential lung cancer biomarkerswith higher sensitivity and specificity were found, with four common biomarkers in both IMAC-Cuand WCX2 chip; the remaining biomarker occurred only in WCX2 chip. Two biomarkers wereup-regulated while three biomarkers were down-regulated in the serum samples from patients withnon-small cell lung cancer. The sensitivities provided by the individual biomarkers were 75%-96.43%and specificities were 75%-100%. Conclusions The preliminary results suggest that serum is acapable resource for detecting specific non-small cell lung cancer biomarkers. SELDI massspectrometry is a useful tool for the detection and identification of new potential biomarker ofnon-small cell lung cancer in serum.