<p>Low-light image enhancement aims to restore high-quality and visually natural results from severely degraded inputs. However, achieving a balance among brightness enhancement, noise suppression, and detail preservation remains highly challenging. Most existing approaches adopt homogeneous strategies or rely on guidance signals carrying limited information, which fail to adaptively handle regions with diverse quality variations. To address this issue, we propose a hierarchical-guidance iterative enhancement network for low-light images. The proposed method formulates the enhancement task as an iterative refinement process in a multi-scale feature space, guided by a perceptual quality map that integrates brightness, noise, and contrast information. Based on this design, we introduce a hierarchical-guidance refinement block, in which the guidance signal functions in two modes: region gating for local enhancement and quality modulation for global context refinement. Moreover, we design a synergistic feed-forward network to effectively integrate local structural details with global semantics. By iteratively applying this block across different levels of the U-shaped network, our approach achieves a unified multi-level enhancement process driven by a high-level prior. Extensive experiments on multiple public benchmarks demonstrate that the proposed method consistently outperforms state-of-the-art techniques in both quantitative metrics and visual quality.</p>