Understanding the effects of panic on crowd dynamics in emergency situations has long been considered necessary for pedestrian evacuation control. In the case of disasters, stampedes caused by panic behaviors occur with high possibility, and pedestrians are crushed or trampled, leading to enormous casualties. To eliminate the computational errors accumulated in the traditional macromodel, a macro-microconversion model based on the SF (social force) model and the AR (Aw-Rascle) model is proposed in this paper. The purpose is to use the crowd parameters of the microscopic model as the input part of the macroscopic model and to combine the advantages of the two models to ensure accuracy and improve calculation performance. The concept of the “pressure term” is defined to measure the panic level of the crowd. In addition, a flowchart of the numerical simulation is designed based on the road network conditions at the trampling site. To validate the conversion model, a numerical simulation is conducted in a case study of the Mecca Hajj stampede in 2015. The simulation results display the whole process of crowd marching and meeting with the dynamic variations of the “pressure term.” The simulation results are compared with the traditional simulation results based on a Gaussian distribution, which verifies that the simulation results obtained by the proposed method are closer to the real situation. Moreover, in this study, a new micromacro transformation method for crowd evaluation dynamics, which can enhance computing speed and execution efficiency, is provided.