Abstract: Flashing green signal for pedestrians at signalized intersections is used to warn remaining pedestrians at crosswalks that their right of way is going to end and they should try to finish crossing before the vehicles at the crossing street are released.A model is proposed to study the behaviors of pedestrians during the flashing green phase.This paper analyzes the effects of the following factors on the crossing behavior of pedestrians,including flashing green signal,sur-rounding pedestrians,boundaries of crosswalks,and right-turning vehicles.Afterwards,a model for crossing behavior is developed based on the social force theory.Coefficients for this model are calibrated by using actual data including specifi-cations of crosswalks and speeds of pedestrians and crossing vehicles.Using the proposed model,the crossing behaviors of pedestrians are simulated.The performance of this model at the macroscopic level is validated by layering and collision avoidance behaviors with conflicting pedestrians and vehicles.By fitting the simulated changing rate of the crossing speed with the observed,the optimization indexes RNLof the first or second half of the crosswalks are 83% and 95% respective-ly,which validate the performance of the model at the microscopic level.