Purpose: Age at diagnosis has been identified as a major determinant of thyroid cancer-specific survival. But the cut-off value for age was controversial. The interaction among gender, age and histologic subtypes needed to be answered. Methods: We identified 59,892 thyroid cancer (TC) patients from the Surveillance, Epidemiology, and End Results (SEER) database. We divided the patients into the following three groups according to age: 20-44 years (young), 45-64 years (middle-aged), and >= 65 years (elderly). Logistic regression model was used to identify factors relating to prognosis in elderly patients. Multivariable Cox regression model identified potential prognostic factors. All statistical tests were two-sided. Results: Elderly patients had significantly worse prognosis than the other two groups, P=0.001. Elderly patients had higher proportion of male gender, advanced tumor grade, follicular subtype and advanced tumor stage. There was no survival difference for elderly patients to receive lobectomy and total thyroidectomy, P=0.852. Cox proportional hazards regression model showed that gender, marital status, histology, tumor grade, tumor size, TNM stage, surgery and radiotherapy were all independent prognostic factors in the multivariable analysis. Male patients with TC had worse prognosis than their female counterparts in differentiated tumor but not in undifferentiated tumor. There were more patients of larger tumor, advanced TNM stage and histologic subtypes in male patients. Conclusions: In conclusion, there were a series of factors contributing to the poor prognosis in elderly patients including clinic-pathologic factors and therapy selection. There was no survival difference for elderly patients to receive lobectomy and total thyroidectomy.