Modern buildings, compared with those in the past, are becoming larger and more complicated. Thus, it has become exceedingly difficult to evacuate and conduct rescue operations when a fire accident occurs. In fact, fire has caused a lot of damage globally. However, there is still a scarcity of methods to predict and prepare for a fire accident. Evacuation simulation was developed to accurately predict and analyze disaster situations and effective evacuation plans in the event of fire. The fire dynamic data obtained using Computational Fluid Dynamics (CFD) applies the evacuation simulation environment via the Fire Hazard Field (FHF) algorithm developed in this research, and the agent with human characteristics algorithm interacts with the FHF and proceeds to evacuate. In this research, the process of mapping fire dynamics data to FHF, interaction between agent and FHF and human characteristics algorithm applied to agent are introduced. In addition, the evacuation simulations regarding various fire situations were conducted and the evacuation performance and damage based on fire characteristics were analyzed. The results of this research indicate not only the importance of interaction between hazardous factors and agents but also the application of human characteristics to agents in the crowd evacuation simulation. This research suggest the possibility of establishing an evacuation strategy regarding fire accident by utilizing simulation method.