Fire safety in urban areas is gaining more attention along with the rapid development of urbanization and the increasing complexity of urban structures. Many fire spread models have been developed to support emergency decision making. However, most of them only investigate on single scenarios (forest fire or house fire) and usually have difficulty in balancing model accuracy and timeliness, which makes it hard to apply these models to large-scale heterogeneous scenarios. In this paper, we developed a fire spread model based on the heterogeneous Cellular Automata model for large-scale complex wildland-urban interface (WUI) areas. The model flexibly integrates a forest fire model and a house fire model based on thermal principles and empirical statistics. A software platform is developed based on the fire spread model correspondingly to validate its performance. Taking a real WUI fire occurred in California as a case study, we conducted comparative simulation experiments based on our software platform and widely used forest fire simulation software FlamMap6. The experimental results show that our model can simulate fire spread process with high efficiency, strong timeliness, and competitive accuracy, providing support on emergency decision-making of large-scale complex fire scenarios.