A street system in a residential area provides many alternate routes between the same origin and destination which vary slightly with respect to length. Because the pedestrians or bicycle users would choose their route reflecting the route characteristics and their preference, many routes between same OD would be used in a probabilistic manner. This paper is to develop the route-choice- model for pedestrians and bicycle users using Dial’s probabilistic multipath assignment mode. In this study, Dial’s model was empirically proved to be superior to the shortest path assignment model through the analysis of work and shopping trips in a real residential area.