With the advancement of global urbanization, urban streetscapes have become a critical part of urban public spaces. As one of the important ways to enhance the living environment, relevant government departments have accelerated the construction of urban streetscapes; however, certain issues, such as emphasis on surface and light protection, still persist. The visual effects of street landscapes hardly meet people's daily and spiritual needs. Therefore, assessing the quality of urban streetscapes has scientific and practical significance for improving urban planning and construction and creating a new intelligent and green city. In this study, the spatial distribution of vehicle interference index, spatial feasibility index, road area index, green visual index, sky visibility index, spatial enclosure index, color richness index, and visual entropy evaluation indices were drawn from the framework of combining deep learning and machine learning to quantify the spatial quality of urban streets by integrating a large number of streetscape images. The spatial quality of street landscape was evaluated according to the ELO scoring mechanism and random forest method, taking the main urban area of Fuzhou City as an example. The results showed that the overall street landscape quality of Fuzhou City was high, and the spatial distribution showed a pattern of concentrated high values in the center and dispersed high values in the periphery. Therefore, based on quantifying the spatial quality of urban streets, strategies, such as the pavement layout of pedestrian and vehicular paths, control of the continuity and permeability of visual landscape elements, and plant color matching, should be regulated. The method proposed in this study can provide novel insights for urban builders to improve the construction of cities.