The geographical system is a complex system including multiple spheres and their interactions. A geographical system model is a coupled human and environmental system model for understanding and predicting changes in patterns and processes of the system. Geographical system models are important tools for supporting sustainable development policy making, and thus is a key scientific research area. During the past decades, geographical system models have been developed to examine global environmental problems such as global change and sustainable development. However, the existing geographical system models have limited ability to simulate and project changes in the complex coupled human- environmental system. This study reviewed the history of development of geographical system models, which is characterized by transitions from single process to multiple process, statistics-based to process-based, static to dynamic, and site level to regional and global scale simulations. We also summarized two important development trends of geographical system models: toward coupled human and natural processes and their interactions, and model- data synthesis. To support policy making for sustainable development in China and globally, developing geographical system models is urgent for the Chinese physical geography research community.