Abstract: To reduce the performance uncertainty caused by traffic accidents in a road network, the partial stochastic user equilibrium was used to describe the status of a road network under the impact of accidents, and a bi-level programming model for reliable network design problem was formulated. In the upper level model, the expected total travel time (ETTT) is adopted as a reliability measure, and the optimization objective is to minimize the weighted sum of the investment costs and the ETTT. The lower level model predicts the traffic flow pattern during accidents by generalizing the partial user equilibrium model to account for not only the partial impact of traffic accidents but also the imperfect information that travelers perceive. A Monte Carlo simulation-based genetic algorithm was provided to solve the bi-level model. The results of a numerical example show that, compared with the existing deterministic method, the proposed model can yield more robust solutions; under three demand levels of 100, 150, 200 veh/h, the average increase of social costs is 18%, but the increase rate of social cost decrease with the increase of demand.