Outbreaks and recalls related to nuts and seeds in the United States have increased recently, and 80% of these recalls are due to Salmonella. The U.S. Food and Drug Administration's Food Safety Modernization Act requires food manufacturers to implement risk-based preventive controls based on scientific and technical evidence. Data are limited on the inactivation of Salmonella during processing of saltwater brined in-shell sunflower seeds. The goal of this research was to validate the adequacy of roasting in controlling Salmonella during the production of sunflower seeds and to assess the resulting risk. Four Salmonella strains were inoculated onto sunflower seeds and processed to simulate commercial manufacturing. Seeds were tumbled and roasted at 2258F (107.28C) and 2758F (1358C) for roasting times from 5 to 45 min. Regression models for Salmonella inactivation and water activity change were developed. The inactivation model predicted a 5-log reduction in Salmonella when sunflower seeds were roasted at 1358C for 19.2 min, with a corresponding water activity of~0.61. Roasted sunflower seeds are typically not saleable at water activities .0.6 due to quality issues. Saleable water activities (0.03 to 0.04) were only achieved when the sunflower seeds were roasted for 45 min at 1358C, which resulted in a .7-log reduction in Salmonella. A quantitative microbial risk assessment based on literature values, expert opinion, and the above-mentioned models was used to predict risk of salmonellosis from sunflower seeds. The quantitative microbial risk assessment model predicted an arithmetic mean probability of illness of 1.45EÀ07 per 28-g serving based on roasting at 1358C for 20 min and an arithmetic mean probability of illness of 5.46EÀ10 per serving based on roasting at 1358C for .45 min (i.e., saleable product process parameters). This study demonstrates that sunflower seeds roasted to saleable parameters should not represent a public health risk from potential presence of Salmonella.