A serum-specific protein 'fingerprint' model was established which is capable of evaluating the effect of chemotherapy (gemcitabine) of pancreatic adenocarcinoma. We used SELDI-TOF-MS coupled with CM10 chips and bioinformatics tools to analyze a total of 45 mouse serum samples from three groups: the healthy control group, the pancreatic cancer model group (orthotopic transplantation model of human pancreatic adenocarcinoma) and the gemcitabine-treated group to establish diagnostic models. As a result, the test set yielded a specificity of 95.0% and a sensitivity of 95.0% for pattern 1, which distinguished pancreatic adenocarcinoma from healthy individuals and a specificity of 95.0% and a sensitivity of 75.0% for pattern 3, which distinguished healthy controls, PC model group and gemcitabine-treated group, as evaluated by leave-one-out cross-validation. We concluded from this study that the SELDI-TOF-MS technique combined with bioinformatics approaches can facilitate evaluating the effect of chemotherapy (gemcitabine) for pancreatic adenocarcinoma and could be used as a potential prognostic monitoring method.