BackgroundWalking plays an important role in residents’ health. After experiencing rapid development, most cities in China have gradually entered the process of urban renewal. Therefore, improving the quality of neighbourhood streets in the old city has become an important issue. The sense of security at night has become an important part of street reconstruction. Therefore, finding out the key environmental factors that affect the security of night-time walking in the streets of the old city has become an important prerequisite for street reconstruction. MethodsFirstly, according to the theoretical explanation of security in the existing literature, the environmental variables that may affect the security of night-time walking are determined. Besides, we chose the typical neighbourhood streets in the old town that need to be improved urgently, and recruited 32 subjects in an on-site experiment to record. Through wearable devices, the galvanic skin response, (GPS), and panoramic video signals generated by the subject during the experiment. The high arousal environment nodes of the subjects were determined by Galvanic skin response GSR signal and GPS signal. According to the semi-structured interview of the subjects, we valenced these high-arousal environment nodes and established a map of each subject’s fear emotional environment nodes. Aggregated into a street walking fear map by kernel density estimation. Finally, we selected the environment nodes with different degrees of fear on the map, established the optimal regression model of night-time walking security through the optimal subset regression method, and finally identified the critical environmental factors. Results and conclusionThe key environmental influences are street transparency, illegal parking level, environmental cleanliness level, light-transmittance intensity form building, CCTV density, and crowd activity intensity. The above key factors suggest that theEyes on the Street (EoS) theory seems to occupy a more important position in the relevant theories of street night-time security research. It also suggests that the privacy of subjects needs to be considered when improving the security of the street at night.