<p>Visual object tracking, particularly in challenging scenarios like low illumination and background clutter, significantly benefits from the complementary nature of visible (RGB) and thermal-infrared (TIR) information. In this paper, we propose an effective RGB-T tracking algorithm that leverages both modalities through a multilevel fusion scheme at the pixel, feature, and decision levels. This approach fully exploits the advantages of each modality, enhancing the robustness of the tracker. Furthermore, we introduce a motion-cue-based visual region correction module that mitigates boundary effects by aligning the visual center based on spatial and temporal motion cues. Extensive experiments on benchmark datasets demonstrate the superiority of our proposed method compared to other RGB-T trackers, validating its effectiveness and robustness in handling diverse challenges in RGB-T tracking tasks.</p>