To decrease the suffering of survivors，improve efficiency of disaster response，and reduce various losses，a leader-follower optimization problem for trans-regional relief reactive scheduling is considered． A bi-level integer programming model that characterizes uncertain relief-supply and demand，varying travel time，equitable principle，multi-level heterogeneity and risk acceptability of survivors，vertical intergovernmental relations of beneficiaries，split demand，multiple supply and demand points，and multi-mode transportation is formulated． The objectives are to minimize total weighted travel time at the upper level and maximize total weighted survivors’perceived satisfaction at the lower level． Then，a primal-dual algorithm based on complementary slackness constraints is designed by analyzing the characteristics of this mathematical model． Finally，a case study from Wenchuan earthquake is presented to illustrate the proposed model and solution strategies．