2015, Korean Society of Civil Engineers and Springer-Verlag Berlin Heidelberg. This paper investigates pedestrians choices between escalator and stairway inside urban rail transit stations during peak hours. After the review of mode choice models, a logit model based on the utility maximization theory is selected. Binary logit models for the passenger flow distribution are established with walking time, walking distance, gender and age of the passenger as the main influencing factors. The influences of these factors are quantitatively analyzed through data collected from three rail transit stations with different arrangement forms of vertical walking facilities in Nanjing, China. The developed models are verified to produce a high accuracy in predicting passengers choices of vertical walking facilities. With developed models, sensitivity analyses for selecting escalator for upward and downward pedestrians in the three types of stations are also conducted and compared to further reveal the differences in pedestrians choices of using vertical walking facilities. The modeling procedure and results can provide a theoretical reference for operation decisions, such as the planning of vertical walking facilities and passenger flow organization in busy rail transit stations.