OBJECTIVE To identify the predictors of cancer-specific mortality of penile squamous cell carcinoma (PSCC) using a population-based database. METHODS Using data from the National Cancer Institute's Surveillance, Epidemiology, and End Results registry, we performed a time-to-event analysis to determine which clinical parameters were useful in predicting cancer-specific mortality. RESULTS Our cohort consisted of 2515 cases of PSCC diagnosed from 1973 to 2007. The patients were divided into 2 groups: primary tumors of the prepuce (n = 722) and primary tumors of the glans, body, and overlapping lesions of the skin (n = 1793). The median follow-up for the cohort was 39 months (range 1-411). Compared with tumors of the prepuce, tumors of the body (hazard ratio 1.61, 95% confidence interval 1.00-2.60, P = .05) and overlapping tumors of the skin (hazard ratio 1.79, 95% confidence interval 1.13-2.83, P = .01) had a greater risk of cancer-specific mortality, even when controlling for age, Surveillance, Epidemiology, and End Results stage, and tumor grade. Furthermore, the disease-specific 10-year survival rate of those with preputial tumors was 89.4% compared with 78.7% for the other 3 groups combined (P < .0001). CONCLUSION Anatomic site-specific disparities for PSCC survival appear to exist. Patients diagnosed with PSCC of the prepuce have greater overall long-term disease-specific survival than patients with primary tumors elsewhere.