Infrared small target detection (IRSTD) plays a critical role in a wide range of applications, including precision strike, early warning systems, forest fire monitoring, and autonomous driving. However, it remains a challenging task due to low contrast, small target size, blurred boundaries, and complex background clutter. To address these issues, we propose a novel Tri-guided Hybrid Attention Network (TriHANet) that jointly exploits multi-scale structure awareness, adaptive channel selection, and spatial consistency modeling for robust and accurate detection. Specifically, we design a Tri-guided Hybrid Attention Module (TirHA) that integrates a Multi-Scale Perception Module (MSPM), an Adaptive Top-K Channel Self-Attention (ATCSA), and a Body-Edge Spatial Attention (BESA) mechanism. This module enables the network to extract structural features at multiple scales, dynamically suppress redundant channels, and enhance spatial responses along both the target body and contour. Furthermore, we introduce a Dynamic Alignment Feature Fusion (DAFF) module in the decoding stage to align multi-scale semantics and reinforce boundary consistency through hierarchical guidance. Extensive experiments on two benchmark datasets, SIRST and IRSTD-1K, demonstrate that our TriHANet significantly outperforms state-of-the-art methods in both quantitative metrics and qualitative visual performance. The results validate the effectiveness and generalization capability of the proposed approach in complex infrared imaging scenarios.