This study focuses on network reconstruction planning after natural disaster that caused large-scale transportation network to be destroyed. The natural disasters such as typhoon or earthquake usually follow the large-scale transportation network destruction. How to repair the transportation network’s basic capability in a timely manner after the occurrence of natural disaster is very important issue. The efficiency of network recovery significantly affects the effectiveness of disaster victim emergency evacuation and rescue. In this research, the emergency network reconstruction problem for natural disasters is formulated as a bi-level programming network design model. The upper level model is a decision of minimized system cost under limited network reconstruction resources. The lower level model is one of the constraints with the upper level problem that considers users equilibrium route choice behaviors under a fixed trips’ demand constraint. In solving the problem, the variational inequality sensitivity analysis method, generalized inverse of matrix approach, and gradient projection method are adopted to develop the solution a algorithms for the bi-level model. The numerical test results indicated that the Stackelberg equilibrium solution of the emergency network reconstruction be-level programming model do exist. More significantly, the optimal network reconstruction planning, disaster victim emergency evacuation and rescue route planning can be achieved.