The environmental assessment of bicycle travel in rail stations catchment areas is critical for promoting the efficiency and quality of bicycle feeder trips. Based on multi-source data from Baidu Street View, road network, and rail stations, the comfort, safety, convenience, and accessibility of bicycle riding environment evaluation indexes were selected. The DeepLabV3 + model and sDNA analysis were used to assess the bicycle travel environment of various types of roadways within the 1 500 m range influence region of the rail station, with Nanjing City serving as an example for empirical analysis. The results show that there are spatial differences in bicycle travel environment in the catchment area of the rail stations. The traffic-type roads have high travel accessibility and high sky visibility, but low bicycle access convenience. Although the accessibility of life roads is low, the convenience of bicycle access and green vision rate are high. Comprehensive roads indicators are relatively coordinated.