Traumatic pelvic injury is one of the most dangerous injuries because it is often associated with severe hemorrhage as well as serious complications. It is therefore vital to provide immediate medical treatment to increase the survival rate of pelvic injury patients. However, it is often difficult to make treatment decisions, as cases are complex and display similar patterns. It has been suggested that the use of computer aided decision-making in a trauma support system is the most efficient way to reduce the cost of trauma care. In our previous work, we found that creating rules using all available variables results in lower accuracy than when using only significant variables. This is because less relevant attributes and/or less reliable attributes with regards to the means of measurement can result in random correlation that is clinically meaningless. Based on this knowledge, we designed an efficient computer assisted trauma decision making system for traumatic pelvic injuries using a machine learning algor thm. More specifically, a rule-based system was designed to create a reliable method of making predictions/recommendations on the status and exact outcome – i.e. home or rehabilitation - of pelvic trauma patients using a nonlinear regression and classification (CART) method. The resulting computer aided system can aid physicians in making rapid and accurate decisions. Three machine learning algorithms were compared to evaluate the proposed method.