Around 1.3 million people die in car accident every year in the world. In Malaysia, car accident has become top ten death cause as the number of road users is increasing. To counter the increasing trend of traffic faulty, Advanced Driver Assistance System (ADAS) is actively investigated recently to assist a driver in the car control. Therefore, traffic sign recognition is one of the important parts in the ADAS. It helps a driver to gain better awareness of the road signs, and eventually alert the driver of the possible rules violation such as the speed limit around the school area. In this project, Speeded_Up Robust Feature (SURF) extraction is used to reduce the dimensionality of the original input traffic sign image. The reduced features are then fed into Radial Basis Function Neural Network (RBFNN) for traffic sign classification. As a result, it improves the recognition rate and shortens the computation time of the traffic sign recognition system. Along with the experiments, Malaysia trafiic sign database is used as the input for the traffic sign recognition system for performance evaluation.