Abstract: In the process of clustering analysis of traffic parameter data, the traditional fuzzy C-means (FCM) algorithm suffers from problems such as high randomness in the initialization process of clustering centers, which can cause fluctuations in clustering centers and fall into local optimal solutions. We propose an FCM optimization method based on Improved Osprey Optimization Algorithm (IOOA) to address these shortcomings, and apply it to achieving accurate classification and discrimination of traffic operation state. the proposed IOOA-FCM uses Logistic-Chebyshev hybrid chaotic mapping to enhance population diversity, and designs a dynamic mixed parasitic symbiosis strategy to balance exploration and development capabilities, combined with a memory-enhanced reverse learning mechanism to improve global optimization performance. We use standard test functions to compare IOOA-FCM with other advanced algorithms, including OOA, SSA, GWO, and MFO. The results show that the proposed IOOA-FCM performs better in terms of convergence speed and accuracy than other algorithms. We then validate IOOA-FCM on the PeMSD8 public traffic dataset, and the experimental results show that compared to the comparative algorithms, the proposed IOOA-FCM algorithm can more quickly and accurately classify the traffic operation state.