Abstract: In order to make traffic assignment be in more accordance with travelers' actual behavior, a continuous function for path prospect was presented based on the cumulative prospect theory. A stochastic user equilibrium model and its equivalent variational inequality were formulated. This model can captures three kinds of uncertainty simultaneously, namely traffic system's uncertainty and errors in travelers' perceptions and modeler's observation. Some properties of the proposed model were discussed, and a corresponding algorithm and numerical example were presented. The research result indicates that with the increasing of travel reliability parameter or risk attitude parameter between interval 0 and 1, travelers will be more sensitive to risk and inclined to choose routes with less travel time variability, while travelers' choices will be stabilized in the event of small perception errors or a high degree of uncertainty. Furthermore, the loss aversion coefficient has a little influence on equilibrium traffic flow.