As European cities age, the challenge extends beyond ensuring older adults can reach essential services to understanding how urban environments support their movement, rest and cognitive engagement in public space. The 15 min city framework emphasizes access to services but often overlooks environmental comfort, spatial legibility and how older adults actually use public space. Existing accessibility metrics systematically mispredict elderly spatial engagement because they ignore micro-scale environmental and configurational factors. This study addresses this gap by developing a grid-based urban walking comfortability model for the neighbourhood of Alvalade in Lisbon. The analysis is conducted using a 100 m × 100 m geospatial grid to capture fine-scale variations in the urban fabric. The work proceeds via three methodologies: ‘Functional Accessibility Overlays’, ‘Environmental and Cognitive Layer evaluation’ and finally ‘Behavioural Validation’ by Computer Vision. All indicators are standardized and systematically integrated into a weighted Age-Friendly Comfort Walkability Index. Our findings suggest that proximity alone does not ensure engagement. This study proposes and empirically validates a calibrated spatial performance model. By linking infrastructure, spatial configuration and behavioural evidence, the proposed framework offers a practical tool for planners with an actionable, data-driven tool to embed holistic environmental comfort within sustainable neighbourhood strategies also fostering equitable and age-friendly urban environments.