Out-of-position (OOP) occupants, especially smaller females and children, are quite vulnerable to injuries caused by deploying airbags. This study is aimed at investigating the effects of various design parameters of a frontal airbag on OOP occupants' resulting injury levels. The study also investigates the predictive capability of the occupant's injury potential using a regression model developed from the response surface methodology. For this, the OOP occupant is represented using a MADYMO 3-year old child dummy. The objective measurements identified as the main injury causation parameters include the fabric permeability, friction coefficient between the occupant and the airbag, and the occupant's initial position. Due to the complexity of the airbag model, small iterations are performed until the numerical model is found to be in reasonable agreement with another study available in the literature. The first part of the work is focused on analyzing the effect of a single parameter on the injury responses of the 3 year old child. The second part of the work covers the development of a set of regression model to predict the combined effects of various parameters on the injury responses of the occupant. It is shown that the regression model is fairly accurate and sufficient in predicting various injury levels to OOP occupants from a deploying airbag.