Aiming at the interaction process between pedestrians and autonomous vehicles, the work aims to propose an exploratory decision-making model of pedestrians crossing the road in front of autonomous vehicles from the perspective of pedestrians. Firstly, the pedestrian crossing scene was analyzed and defined. According to the pedestrian crossing intention and traffic intention, two typical pedestrian crossing scenes were summarized. Then, the influencing factors of pedestrians crossing the road in front of autonomous vehicles were classified by latent variable analysis. The external human-machine interface was designed, and the influencing factors and measurement scale were constructed. Finally, the orderly Logistic regression method was used to analyze the influencing factors of pedestrian crossing behavior on pedestrian crossing intention and traffic intention, and a decision-making model of pedestrian crossing the road in front of autonomous vehicles was constructed. The correlation and internal relationship between the influencing factors of pedestrian crossing behavior and their crossing decisions were analyzed quantitatively and the decisive interaction mode between pedestrians and vehicles and the changing process of each influencing factor in the process were put forward. A pedestrian crossing behavior decision-making model considering eHMI is proposed in this work, which divides the pedestrian crossing decision-making process into three stages, and summarizes the decisive change curve of vehicle behavior cues and eHMI cues, as well as the interaction process and key nodes between pedestrians and autonomous vehicles.