Abstract: Infrared target tracking often encounters challenges such as target occlusion, motion blur, and background clutter, which significantly degrade tracking performance. To address these issues, this paper proposes a correlation filter-based tracking algorithm incorporating adaptive regularization and sub-quadratic temporal context response. First, we design a structure-guided non-local means denoising method to improve the quality and discriminability of infrared images. Then, we introduce an adaptive spatial regularization mechanism based on dynamic weights and Gaussian modeling to suppress background interference and enhance tracking robustness. Finally, we develop a sub-quadratic temporal context response regularization that fully exploits historical response information, reduces abnormal response effects during filter updates, and strengthens the model’s accuracy and robustness in complex scenes. Experimental results show that our TCRSCF algorithm achieves the highest success rate and precision on the LSOTB-TIR dataset, improving the baseline by 4.6% and 5.9%, respectively. On the PTB-TIR dataset, it further increases the success rate and precision by 3.3% and 5.1%. Extensive experiments on the VOT-TIR2015 and VOT-TIR2017 datasets also confirm that our method outperforms existing state-of-the-art algorithms in infrared target tracking.