This study presents mathematical surrogate models, derived from finite element kinematic response data, to predict car crash-induced occupant head and neck injury risk for a broad range of impact velocities (10 – 45 mph), impact locations, and angles of impact (-45° to 45°). The development of these models allowed for wide-scale injury prediction while significantly reducing the overall required number of impact test cases. From these, increases in both the impact velocity and the impact’s locational proximity to the occupant were determined to result in the greatest head and neck injury risks. Additionally, strong interactions between the impact orientation variables (location and angle) produced significant changes in the head injury risk, while the neck injury risk was relatively insensitive to these interactions; likely due to the uniaxiality of the current standard neck injury risk metrics. Overall, this methodology showed potential for future applications in wide-scale injury prediction or vehicular design optimization.