This study tested the hypothesis that prediagnostic soy intake was inversely associated with all-cause and breast cancer-specific mortality. The analyses included 3842 women in the Multiethnic Cohort (MEC) study of African Americans, Native Hawaiians, Japanese Americans, Latinos, and Caucasians, who completed a quantitative food frequency questionnaire, aged ≥50 yr at cohort entry, and diagnosed with primary invasive breast cancer following cohort entry (1993-2007). Hazard ratios (HR) and 95% confidence intervals (CI) were estimated from Cox proportional hazards regression with adjustment for known clinical and lifestyle factors. During a mean follow-up after diagnosis of invasive breast cancer of 6.2 ± 3.8 yr, there were 804 deaths including 376 breast cancer-specific deaths. The HR (95%CI) for all-cause and breast cancer-specific morality comparing the highest versus lowest tertiles were 1.03 (0.81-1.33) and 1.03 (0.71-1.50) for soy products and 0.99 (0.82-1.20) and 0.95 (0.71-1.28) for total isoflavones, respectively (Ptrend > 0.60 for all). There was limited evidence of differences by hormone receptor status, tumor stage, or ethnic group. Prediagnostic soy intake was unrelated to mortality in postmenopausal women. Our findings are consistent with the literature that soy consumption does not adversely affect breast cancer survival in women.