Abstract: Based on passenger distribution characteristics analyzed in subway station platform,electric theory was introduced and intensity function of waiting site was set up.The passenger prediction model for waiting sites in subway stations was presented.To estimate the parameter,an objective function of minimum squared deviation between actual passenger distributions and the predicted in waiting areas was put forward.In order to estimate this problem,the advance and retreat method of fixed range was improved.Then,the relationship between this parameter and the passenger distribution was revealed.Finally,passenger distribution data of Haidianhuangzhuang Station was adopted to validate the model.Maximum error between the actual passenger distributions and the predicted was 6 passengers.Most of the errors were between 0~4 passengers,which could meet the actual requirements.