The effective identification and categorization of emergency vehicles is essential for improving traffic management systems, which results in quicker response times and safer travel for emergency responders. We introduce a new approach that merges Histogram of Oriented Gradients (HOG) feature extraction with a Support Vector Machine (SVM) classifier for real-time separation of emergency and non-emergency vehicles. Optical Character Recognition (OCR) technology extracts the text from vehicle markings after classifying the vehicle type. Fuzzy matching techniques are used to match extracted text with predefined vehicle types to improve vehicle type identification accuracy for categories like ambulances and fire trucks. Research results show that this combined system produces high classification accuracy and dependable detection of emergency vehicles, which makes it an effective solution for real-time urban traffic management applications. Emergency vehicle detection and traffic flow management can experience substantial enhancements through the implementation of the proposed method.