Laryngeal carcinoma is the most common malignancy among head and neck tu-mors. The purpose of this study is to find biomarkers for laryngeal carcinoma in patient blood serum using the Surface Enhanced Laser Desorption/Ionization (SELDI) technique. Serum samples from 33 laryngeal carcinoma (12 cases of glottis, 18 of supraglottis and 3 of subglottis) patients and 31 age- and sex-matched healthy people were analyzed by SELDI-TOF on a Pro-teinChip reader, PBSII-C. Protein profiles were generated using WCX2 protein chips. Protein peak clustering and classification analyses were performed utilizing the Biomarker Wizard and Biomarker Pattern software packages, respectively. The results showed that sixteen peaks had significant difference between laryngeal cancer patients and healthy group, eight of which were up-regulated in the patient samples, and the others were down-regulated. Two protein peaks 8153 Da and 2035 Da were automatically chosen for the system training and development of a classification tree. The analysis yielded a correct percentage of 96.9% for patients and 96.7% for control. The results suggest that serum is a useful resource for the detection of specific bio-markers for laryngeal carcinoma. Proteinchip Array System was a useful tool for a high throughput screening of large-sized serum samples to discover potential biomarkers for carci-noma. Lilyngeal carcinoma is the most common malignancy among head and neck tu-mors. The purpose of this study is to find biomarkers for laryngeal carcinoma in patient blood serum using the Surface Enhanced Laser Desorption / Ionization (SELDI) technique. Serum samples from 33 laryngeal carcinoma (12 cases of glottis, 18 of supraglottis and 3 of subglottis) patients and 31 age- and sex-matched healthy people were analyzed by SELDI-TOF on a Pro-teinChip reader, PBSII-C. Protein profiles were generated using WCX2 protein chips . Protein peak clustering and classification analyzes were performed utilizing the Biomarker Wizard and Biomarker Pattern software packages, respectively. The results was that sixteen peaks had significant difference between laryngeal cancer patients and healthy group, eight of which were up-regulated in the patient samples, and the others were down-regulated. Two protein peaks 8153 Da and 2035 Da were automatically chosen for the system training and development of a cla The analysis suggests that a serum is a useful resource for the detection of specific bio-markers for laryngeal carcinoma. Proteinchip Array System was a useful tool for a high throughput screening of large-sized serum samples to discover potential biomarkers for carci-noma.