Wildfires pose a big threat to human life and property safety. Previous studies on wildfire risk management focused mainly on understanding wildfire behaviour through computer simulations. Effective wildfire risk management also largely depends on the evacuation performance success. Computational tools tend to be the best approach for simulating wildfire evacuation emergencies as well. Therefore, this study proposes a comprehensive simulation framework that integrates Agent-based Modelling (ABM) and Geographical Information Systems (GIS) to efficiently simulate both human behaviour and transportation crowds. In particular, ABM bridges the technical gap between GIS and a multi-agent system (MAS) for simulation design efficiently and effectively. To study the evacuation performance (measured by the number of agents being sheltered or refused to evacuate), our modelling solution enables altering relevant model parameters in wildfire evacuation scenarios. The simulation outputs, as a result, can be used to evaluate the influential factors and further assist in effective evacuation planning. In particular, the following tasks are performed: (1) simulation of the influence of transportation crowds on evacuation performance; (2) evaluation of the effectiveness of public notification on evacuation success; and (3) comparison of the differences among various transportation means as well as their performances during a wildfire emergency. A case study is conducted to verify the simulation framework proposed in this study. The simulation outputs showed that the transportation crowds negatively impact on the evacuation performance, while public notification can enhance resident risk perception, thus assist in the evacuation efficiency. Finally, public vehicles such as public buses have the highest evacuation efficiency compared to other transportation means tested in this study. Keywords ?