By comparing the clinicopathological features and survivals between the resected and non-resected intraductal papillary mucinous neoplasms (IPMNs) of the pancreas, this study tried to clarify the natural history of IPMNs, to provide a strategy for treatment, and to determine the justification of not performing resection for some patients. A total of 57 patients with IPMN, including 39 resected and 18 non-resected IPMNs, were recruited for study. Data on demographics, clinical presentations, diagnostic work-up, treatment modality, clinical course, and outcomes were evaluated and compared between the resected and non-resected IPMNs. The most common clinical presentation was abdominal pain (57% in total IPMNs, 67% in resected, 33% in non-resected), followed by body weight loss (32% in total IPMNs, 33% in resected, 28% in non-resected). The sensitivity in the diagnosis of IPMN was highest by magnetic resonance cholangiopancreatography (MRCP) (88%), followed by endoscopic retrograde cholangiopancreatography (ERCP) (68%), and computed tomography scan (CT scan) (42%) and sonography (10%). The median survival was 21.5 months for patients with resected IPMNs, ranging from 2 to 124 months, and 14 months in non-resected IPMN patients, ranging from 5.5 to 70 months. There is no significant survival difference between the resected and non-resected groups, with a 5-year survival of 69.8% in resected IPMNs and 59.8% in non-resected IPMNs, P = 0.347. The survival outcome of the unresectable non-resected IPMNs was much inferior to the resected IPMNs, P = 0.002 and resectable non-resected IPMNs, P = 0.001. Thus, the prime prognostic factor in predicting the survival outcome of IPMNs is resectability, instead of resection itself. Long-term survival could also be expected in resectable IPMNs without resection. No resection for the IPMN may be justified for patients with high surgical risks, especially for those who are asymptomatic and very aged.