Abstract: Objective: To search for potential serum biomarkers for gastric cancer screening using serum proteomic based technologies. Methods:Serum samples from 16healthy volunteers (N group),28gastric cancer patients (T group), and 18controls were collected. The differentially expressed peaks of serum proteins were found and selected by the WCX magnetic bead assisted MALDI-TOF-MS technique and ClinProTool software. These data were then used to build a diag-nostic model for gastric cancer. Results:A total of 25peaks of serum proteins were found to be differentially expressed be -tween the tumor group and the control group ( P<0.01). Two of them (m/z: 5248.49and 5754.25) were selected as the po -tential cancer biomarkers by ClinProTool software (sensitivity 84.61% and73.07% , respectively; specificity100 % and 93.75% , respectively). According to the ANN, we selected6 peaks of protein (4268.05m/z, 5636.53m/z, 5248.49m/z, 2933.15m/z, 1450.13m/z and 1349.4 m/z) to build a diagnostic model for gastric cancer. The sensitivity and specificity of this model were 100 % and 75%, respectively. Conclusion:Specific protein markers of gastric cancer can be detected in se-rum by magnetic bead assisted MALDI-TOF-MS. A classification model of serum proteomics could be created for diagnosis of gastric cancer.