Abstract: To reduce the risk of transportation planning scheme and improve the whole benefit of project selection in different stages,a multi-stage discrete network design model under stochastic OD demand is set up to determine the optimal project scheduling with stochastic bi-level optimization.The proposed model,a combinatorial optimization problem,is solved using Monte Carlo simulation and genetic algorithm.Numerical results on Nguyen Dupuis network indicate that the distribution of funds has a great impact on decision making and the system performance can be improved with more funds in early days of the whole planning period.Besides,demand uncertainty and planner's preferences are also crucial to the final program.