Urban walkability depends on more than the physical form of streets. It is also shaped by how pedestrians see, interpret, and respond to their surroundings. Street width, greenery, enclosure, signage, and pedestrian facilities may all affect walking experience. Yet their influence is not only direct. These features first attract, guide, or disperse visual attention. This visual process may then shape how people evaluate safety, comfort, preference, and ease of movement. In recent years, eye-tracking has become an important method for studying pedestrian experience. It allows researchers to record where people look, how long they fixate, and how visual attention changes across different street environments. However, the evidence remains fragmented. Most studies examine specific design features or isolated visual indicators. Few studies explain how visual attention connects street design with walkability-related perceptions and behaviours. As a result, the mediating role of visual attention is still unclear. To address this gap, this study systematically reviews eye-tracking research on urban walkability. The review follows the PRISMA 2020 guidelines. Studies were searched in the Web of Science and ScienceDirect databases. The search covered publications from 2016 to 2026. After screening, 17 studies met the inclusion criteria and were included in the final analysis. The selected studies were examined through four analytical categories. These included country and research context, experimental setting, environmental and design variables, and eye-tracking indicators. The findings were then synthesized qualitatively to identify recurring patterns across the literature. Four main themes emerged from the review. These were greenery and natural elements, spatial openness and scale, walkability infrastructure, and microscale elements and landmarks. Across the reviewed studies, vegetation, spatial coherence, and high-quality pedestrian infrastructure were often linked to more stable and focused visual attention. These features were also associated with higher perceived safety, stronger environmental preference, and better restorative experience. By contrast, visual clutter, excessive information, and spatial discontinuity tended to disperse attention and increase cognitive load. Overall, the findings suggest that visual attention can act as a mediator between street-level environmental characteristics and walkability outcomes. In other words, pedestrians do not simply respond to urban form as a physical condition. They first process it visually. This visual processing influences how they perceive, evaluate, and use street environments. The review therefore provides a basis for future causal models and supports more human-centered approaches to street design.