<p>Infrared small target detection has significant applications in military and security fields. However, because of the small size of the targets and low signal-to-noise ratio, traditional detection methods have limited effectiveness. Existing deep learning methods face challenges in infrared small target detection, such as poor adaptability to complex backgrounds, strong data dependencies, and limited generalization capability. To address these limitations, this paper attempts to enhance the performance of the Yolo algorithm in infrared small target detection based on contrast enhancement strategies. First, a DCDC method is proposed that highlights the differences between the target and the background through contrast enhancement preprocessing. Second, the Attn-MLCL module is introduced, which utilizes multiscale contrast learning to generate richer feature representations. Finally, a composite loss function that integrates target region contrast optimization and gradient feature enhancement is incorporated to further improve the model localization accuracy for small targets. Experimental results show that the proposed method significantly outperforms the traditional Yolo algorithm in infrared small target detection tasks, with notable improvements in detection accuracy and recall rate.</p>