緒論：身體活動量不足是高齡者健康的行為危險因子。「步行」是最適合高齡者身體活動的形式之一，而「環境」又為影響高齡者從事步行行為之重要決定因子。本文將以生態學架構為基礎，進一步探討影響高齡者步行的重要環境因素，包括客觀的「建成環境因素」與主觀的「自覺環境因素」。方法：本研究欲透過系統性文獻回顧，歸結有關客觀建成環境、主觀自覺環境因素與高齡者主、客觀步行行為的未來研究方向與建議。結果：符合納入條件的19篇研究中，18篇文章為量性研究，且均為橫斷性設計、另1篇為質性研究。此外，多數研究使用問卷測量步行行為，只有少數採用客觀儀器測量。整體而言，納入研究中的環境因素變項以主客觀的「環境宜走性(walkability)」為核心概念，研究結果顯示：密度、土地使用多樣性、安全、友善環境等面向與高齡者的不同步行行為類型呈正向關聯性。結論：未來仍需要更多的縱貫性世代追蹤或準實驗研究設計來進一步了解環境與高齡者步行之因果關係。其次，自我選擇(self-selection)因素需納入控制變項，以釐清環境與高齡者步行行為的關聯性。第三、步行行為測量工具，宜兼採主觀與客觀測量以提升準確性。第四，各研究對於高齡者的居家緩衝區(buffer zone)並不一致，難以界定高齡者近鄰環境之範圍。第五，考量到環境因素測量的限制，Walk score®網站為新興客觀宜走性測量工具，然而，其於亞洲國家之效度仍需檢驗。最後，有鑑於亞洲國家在文化、人口密度及環境建構與西方國家的差異，需進一步提出適合亞洲國家的宜走性指標。

Introduction: Physical inactivity is a behavioral risk factor of elderly's health. Walking behavior has been recognized as a suggested form of physical activity for older adults. Evidence shows that environment is a determinants of older adults' walking behavior. According to Ecological model, this study aims to review the literature between environment and older adults' walking, including objectively-measured built environment and perceived environment. Methods: This study using systematic review to better understand the associations of objectively-measured and subjectively environmental factors with older adults' objective and subjective walking behavior, in order to provide future suggestions. Results: Among 19 included studies, 17 were quantitative studies using cross-sectional design. Overall, included studies used walkability as core concept of environmental factors. The results showed that density, safety, land use mix and friendly environment is positively associated with different types of walking behavior among older adults. Besides, more studies using questionnaire to assess older adults' walking behavior, fewer studies have used objective measures. Conclusion: First, more qualitative studies are still needed to better understand the perceived barriers and facilitators of walking behavior among older adults. In addition, self-selection should be included as a key confounder to clarify the influence of environment on walking behavior. Moreover, prospective or quasi experimental design not only can minimize the self-selection bias (active individuals tend to choose neighborhoods conductive to physical activity), as well as improving evidence related to causality. Third, studies can be beneficial to utilize both objective and subjective measures of walking behavior for accuracy. Fourth, the definition of buffer zone varied in different studies and thus it is difficult to define the neighborhood environment of older adults. Fifth, for considering the limitation of environmental measures, Walk score^® website is a novel tool for measuring walkability. However, the validity of Walk score^® website should be further tested in Asian countries. Finally, considering the different culture, residential density and infrastructure between Asian and Western Countries, it is important to further provide the indicators of Walkability for Asian countries.