Pancreatic cystic neoplasms (PCNs) are a group of entities with distinct risks and various treatments. Identification of the PCN patients at risk is thus critical. A correct diagnosis is the key to select high-risk patients. However, the misdiagnosis rate is extremely high even computer tomography, magnetic resonance imaging, and endoscopic ultrasonography were applied. Current approaches for differential diagnosis and identifying high-risk patients in certain types of PCNs are not powerful enough to make a clinical acceptable accuracy of diagnosis. The approaches mainly rely on imaging and tumor marker test. We here summarized the current approaches, and reviewed novel approaches under development. For instance, cyst fluid test of glucose or vascular endothelial growth factor A shows the best performance in identifying mucinous cystic neoplasms or serous cystic neoplasms. Multidisciplinary team (MDT) discussion is another way to improve the accuracy of diagnosis. Combination of MDT with validated novel approaches with high sensitivity and specificity is the best way to select truly high-risk patients with PCNs.