Objective To construct and verify a diagnostic model using proteomic analysis of serum samples for identifying gastric precancerous lesions and gastric cancer (GC). Methods The serum samples from 25 patients with gastric precancerous lesions (chronic atrophic gastritis with mild to moderate dysplasia), 25 GC patients and 25 healthy controls were analyzed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Spectral peaks with significant difference among the groups were identified and used as a diagnostic model for detecting gastric precancerous lesions and GC. The serum peptide map model was validated using an independent sample set including 15 healthy volunteers, 15 precancerous and 15 GC patients. Results The spectral peaks for the peptides with mass-to-charge (m/z) values of 1741 and 4210 were the most significantly different among the three groups. The sensitivity of this diagnostic model for detecting healthy controls, patients with gastric precancerous lesions and patients with GC was 80.0% (12/15), 66.7% (10/15) and 66.7% (10/15) respectively, while the specificity was 66.7% (20/30), 73.3% (22/30) and 73.3% (22/30), respectively. Conclusion Our diagnostic model is useful for diagnosing gastric precancerous lesions and GC.