<p>Reducing traffic congestion necessitates accurate and reliable traffic state discrimination. However, the inherent randomness and nonlinearity of traffic flow present significant challenges to precise traffic state discrimination. To address these challenges, this study proposes a novel hybrid model that integrates several advanced algorithms: the Fuzzy C-means (FCM) clustering algorithm, the Multistrategy Dung Beetle Optimization (MSDBO) algorithm, the Adaptive Boosting (AdaBoost) algorithm, and the Support Vector Machine (SVM). First, the FCM clustering algorithm classifies traffic flow data into five distinct states. Second, the SVM penalty factor and kernel parameters are optimized using the MSDBO algorithm. Next, the AdaBoost algorithm enhances the SVM model by incorporating it as a weak classifier. Finally, the optimized SVM model is employed to establish a comprehensive traffic state discrimination framework. The proposed MSDBO algorithm and the hybrid AdaBoost-MSDBO-SVM model are experimentally validated. First, the performance of the proposed MSDBO algorithm is compared with that of five other methods using eighteen test functions. The results show that the MSDBO algorithm proposed in this study achieves higher solution accuracy and faster convergence. Second, the proposed AdaBoost-MSDBO-SVM model is compared with nine other models, and the results demonstrate that the proposed models outperform all others.</p>