<p>Multi-object tracking (MOT) is widely applied in unmanned vehicles, military reconnaissance, and video surveillance. However, practical tracking scenarios often involve challenges such as partial or prolonged occlusion and object deformation, leading to object loss, incorrect data association, and trajectory fragmentation, all of which compromise tracking accuracy and robustness. To address these limitations, we propose a multi-object tracking algorithm based on the stage-wise association strategy enhanced by weak cues. Our approach introduces stage-wise association, classifying trajectories according to occlusion severity and establishing a prioritized association sequence. By integrating speed-direction consistency and confidence discrepancy as weak cues, we develop customized association strategies for distinct tracking phases, significantly improving association precision. Furthermore, we integrate Generalized IoU (GIoU) to more accurately characterize spatiotemporal relationships between trajectories and detection boxes. Finally, we design an appearance feature update module to mitigate tracking drift caused by object deformation during motion. Experimental results demonstrate that our approach outperforms existing approaches, achieving state-of-the-art performance.</p>