Abstract: Urban streets, as the core carriers of daily travel and social interaction for older adults, directly influence their mobility and quality of life through spatial quality. Taking 3,457 streets in the central urban area of Lanzhou as the research object, this study integrates Baidu Street View images and POI data, employs the Mask2Former model to extract street view semantic features, and constructs an age-friendliness evaluation model of street spaces from four dimensions: safety, convenience, comfort, and inclusiveness. The spatial morphology, zonal differences, and functional types of street spaces in Lanzhou are analyzed. The results show that: The overall age-friendliness of street spaces in Lanzhou exhibits a significant "high in the core—low in the periphery" gradient pattern, with high-value streets mainly concentrated in Chengguan District and areas with mature built environments. Among the four dimensions, safety performs relatively well, while convenience is the main shortcoming, indicating that most streets in Lanzhou have provided basic support for the passage of older adults. Significant differences in age-friendliness exist among different street types in Lanzhou. Life-service-oriented, pedestrian-friendly, and residential-supporting streets perform better, while traffic-dominated and special-passage streets score lower. The study shows that street age-friendliness depends not only on travel safety and facility provision but also closely relates to environmental comfort and support for staying and social interaction.