Abstract: Road signs classification is becoming one of the most significant technologies in driver assistance systems due to the importance of recognizing speed-limit signs to inform and alert drivers when they have limited conscious awareness. In this study, a robust approach for Speed-Limit Signs (SLS) detection and recognition with different distances and situations is proposed. The proposed system is based on a strategy that consists of three distinct steps. First, input image is preprocessed in two sub-steps; binarization based on HSV color space, then noise removal to eliminate undesired noise. Secondly, the processed image is segmented using projection profile to detect the SLS borders dimensions. Finally, the extracted SLS is identified and recognized to show the maximum authorized speed limit. The proposed approach is experimented on Jordanian SLS samples and the results demonstrate that our proposed system can detect and recognize SLS with an accuracy of 92 and 80%, respectively.