As the global population ages, respiratory health among the elderly has become a key public health concern. Although urban green infrastructure (UGI) has the potential to improve air quality and promote health, research on how its layout patterns influence respiratory health among older adults remains limited. This study focuses on elderly residents aged 60 and above in the central urban area of Shenyang, China, to evaluate the relative importance and interactions of different features affecting respiratory health. We utilized the St. George’s Respiratory Questionnaire (SGRQ) to collect data on respiratory health and employed hierarchical regression and random forest (RF) models to analyze the impact of UGI factors across three spatial scales (300 m, 500 m, and 1000 m). The results indicate that UGI within a 300 m radius of participants’ residences contributes most significantly to respiratory health, with diminishing marginal effects as the spatial scale increases. Green space area (GSA) and the NDVI were identified as the most important factors influencing respiratory health, while green landscape pattern metrics had a greater influence at larger spatial extents. Additionally, a significant nonlinear marginal effect was observed between UGI and respiratory health. These findings provide key insights for health-oriented urban planning and green infrastructure design.