Abstract: With the deepening of intelligent mine construction, lithium batteries have become the core power source for mining electric trackless rubber-tyred vehicles, roadheaders, support transport vehicles, and other equipment due to their advantages such as high energy density and long cycle life. However, the underground environment is complex and variable, with harsh working conditions including high temperature, high humidity, impact, and vibration, which significantly increase the risk of lithium battery thermal runaway and pose a serious threat to mine safety production. This paper provides an in-depth analysis of the causes and evolution mechanisms of thermal runaway in mining lithium batteries. Addressing the shortcomings of existing monitoring technologies, it proposes an integrated safety monitoring and thermal runaway early warning system architecture for mining lithium batteries, which combines multi-parameter sensing, edge computing, fusion algorithms, and cloud-based early warning. The study focuses on an early warning model for thermal runaway based on multi-source information fusion（voltage, current, temperature, gas, etc.）. This research demonstrates that this technical solution can effectively achieve early and accurate warning of thermal runaway, providing important technical support for the safe operation of mining lithium battery-powered equipment.