At present, most emergency scheduling models are constructed on single-layer rescue points, but these points may double or even become multilayer rescue points in actual emergency resource scheduling. Based on nonlinear continuous supply and consumption of emergency resource scheduling, we developed an emergency resource scheduling model with two layers of affected points, i.e., distribution centers and reserve bases, to minimize total cost and reaction times. Based on the characteristics of multi-targets and multi-layers of the emergency resource scheduling model, our improved artificial bee colony algorithm uses opposition-based and comprehensive learning concepts to obtain Pareto-optimal solution sets and then analyzes the number of solutions for a Pareto front and uniformity measure. The simulation results show the model to be reasonable and the algorithm to be effective. Therefore, the proposed algorithm can minimize costs and achieve quicker overall reaction times.