Pancreatic cancer is the most lethal of all the common malignancies. Markers for early detection of this disease are urgently needed. Here, we optimized and applied a proteome analysis of human pancreatic juice to identify biomarkers for pancreatic cancer. Pancreatic juice samples, devoid of blood or bile contamination, were collected from patients with pancreatic cancer (n = 5), benign pancreatic diseases (n = 6), or cholelithiasis (n = 3) during endoscopic retrograde cholangiopancreatography (ERCP). After ultramembrane centrifugation sample preparation, pancreatic juice proteins were separated by 2-DE and identified by MALDI-TOF-MS. A 2-DE dataset of pancreatic juice from patients with cholelithiasis was established, consisting of 76 protein spots representing 22 different proteins. Disease-associated obstruction of the pancreatic duct strongly effected the protein composition of pancreatic juice. Concurrently, pancreatic juice from patients with pancreatic cancer was compared to nonmalignant controls with comparable obstruction of pancreatic ducts. Seven protein spots were identified that consistently appeared at changed levels in pancreatic juice from patients with pancreatic cancer. In conclusion, comparative proteomic analysis of human pancreatic juice can be used to identify biomarkers of pancreatic cancer.