Abstract Background: Prior studies of prostate cancer survivors suggest that smoking might be associated with higher prostate cancer–specific mortality (PCSM) after diagnosis with prostate cancer. However, most of these studies were small, and questions remain regarding this association's strength and whether it persists after adjustment for stage and Gleason score. Methods: This analysis included men diagnosed with nonmetastatic prostate cancer between enrollment in the Cancer Prevention Study II Nutrition Cohort in 1992–1993 and June 2013. Cigarette smoking was self-reported at enrollment and updated in 1997 and every 2 years thereafter. Analyses of pre-diagnosis and post-diagnosis smoking included 9,781 and 9,111 prostate cancer cases, respectively, with vital status follow-up through 2014. Results: There were 672 deaths from prostate cancer in analyses of pre-diagnosis smoking and 554 in analyses of post-diagnosis smoking. In multivariable-adjusted Cox proportional hazards regression models including stage and Gleason score, both current smoking before diagnosis [HR = 1.50; 95% confidence interval (CI), 1.06–2.13] and current smoking after diagnosis (HR = 1.71; 95% CI, 1.09–2.67) were associated with higher PCSM compared to never smoking. Prostate cancer survivors who quit smoking <20 years before diagnosis were also at significantly higher risk of PCSM (HR = 1.29; 95% CI, 1.04–1.61). Conclusions: This large prospective study suggests that current smoking both before and after diagnosis of prostate cancer is associated with higher PCSM, even after accounting for stage and Gleason score. Impact: Our results provide evidence that smoking is a relevant prognostic factor for prostate cancer patients and that prostate cancer may be among the causes of death attributable to smoking. Cancer Epidemiol Biomarkers Prev; 27(6); 665–72. ©2018 AACR.