In this population-based retrospective study, we aimed to investigate the association between age at diagnosis and prognosis of pancreatic cancer (PC) patients using data from the National Cancer Institute's Surveillance, Epidemiology, and the End Results database. Different factors for stratification, like race, sex, year of diagnosis, pathological grade, American Joint Committee on Cancer stage, historic stage, and tumour location, were included to compare the survival rates of patients of different age groups, and the five-year survival rate was calculated. Multivariate analysis using Cox regression was performed to control for confounder bias, and the hazard ratio was calculated. In total, 126,066 patients were enrolled in this study. The five-year PC-specific survival of patients aged 20-40 years was almost three times that of patients aged >40 years. Stratified by race, sex, year of diagnosis, pathological grade, clinical stage, and tumour location, a descending trend of survival was observed with an increase in age. On multivariate analysis, the mortality risk of PC patients aged 40-80 years was twice that of patients aged <40 years; however, patients aged >80 years had a mortality risk three times that of patients aged <40 years. The survival rate of PC patients has improved in the last few decades. Age at diagnosis is a significant and negative prognostic factor for PC, and patients diagnosed at a relatively earlier stage had the best survival.