This paper focuses on establishing a new ranking system using an optimization methodology for ordering the AIS injury codes. Codes are assigned mortality rate values, which can be used to get an idea on which injuries are critical and thus should be prevented and how many lives will be saved as a result of preventing that injury. For better statistical correlation, an injury coding scheme is applied which condenses 551 7- digit AIS injury codes down to 50 unique codes. The crash victim’s injury profile is characterized by the top three injuries which are used to calculate the victim’s probability of fatality. Optimization solvers are used to assign mortality rates to each of the 50 codes. Based on the average deviance value, fatality predictions using the optimized mortality rates show better predictive capability over other schemes, including the Maximum AIS score and the Injury Severity Score. Comparisons are made between results derived from the Crashworthiness Data System (NASS-CDS) and the Crash Injury Research and Engineering Network (CIREN).