Abstract: Single-target tracking as a pivotal task in the field of computer vision, finds extensive applications in various domains such as security surveillance, human-computer interaction, industrial automation, and consumer electronics. With the rapid advancement of computer vision technologies, deep learning-based target tracking algorithms have achieved remarkable results. However, current mainstream methods still confront numerous challenges when dealing with complex scenarios such as rapidly moving targets or brief occlusions. In view of this, this paper proposes a novel real-time single-target tracking algorithm with full-image search based on deep convolutional neural networks. The algorithm integrates a full-image search strategy and a lightweight Siamese single-target tracking model, as well as an information interaction module with a self-attention mechanism. It can significantly enhance the robustness and continuity of target tracking, and is particularly suitable for real-time applications on edge devices. Experimental validations demonstrate that the proposed algorithm maintains superior tracking performance across a range of complex scenarios.