Near real-time spatial prediction of coseismic landslides is pivotal, enabling rapid forecasting of the potential locations of widespread landslides in the immediate aftermath of a severe earthquake. Such predictive capabilities are crucial for optimizing the 72-h ‘golden window’ of opportunity for search and rescue operations, which is critical for survivor recovery. Nevertheless, the forecasting of landslides induced by earthquakes presents formidable challenges, primarily due to the limitations in predictive accuracy and the prevalence of false alarms, which can undermine the effectiveness of emergency response measures. This chapter suggests innovative, sophisticated, and interpretable models developed for the near real-time prediction of coseismic landslides, with a focus on the Tibetan Plateau, a well-known seismically-active region, as well as extending its application to a global scale. The models aim to provide precise and timely predictions to aid in disaster mitigation and response efforts in areas prone to seismic activities. Furthermore, this chapter delves into the pivotal factors that dictate the spatial distribution of earthquake-triggered landslides, revealing the underlying patterns that contribute to this phenomenon. This chapter presents an in-depth examination of the distribution patterns of coseismic landslides across the globe, offering insights that can enhance our understanding and predictive capabilities regarding these geological events.