As an intuitive indicator for observing the green environment of roads, street green visibility represents the level of green landscape seen from the perspective of pedestrians, and it is one of the most critical indicators for studying the quality of road greening in recent years. This study extracts the green visibility of some streets in Chengguan District, Lanzhou City, by applying technologically advanced Baidu street view data, combined with innovative deep learning methods, to deeply analyze the key elements of street greening quality, assess the level of street greening, and put forward relevant development suggestions. The study shows that the proportion of vegetation elements in Baidu street view pictures can reflect the perception of street greenness from the pedestrians’ point of view to a certain extent; the level of street greenness in the 12 streets within the study area of Chengguan District, Lanzhou City, is generally better; street greenness has a strong positive correlation with the crown width of street trees, a positive correlation with the height of the tree, a negative correlation with the spacing between the trees, and a significant positive correlation with the level of planting. This study expands new data sources for street green visibility, provides quantitative support for accurately improving the greening quality of roads in Chengguan District, contributes to scientific decision-making, and further promotes the refinement of landscaping construction.