Abstract: Big data technology provides an innovative path for the transformation of mine safety management from passive response to active prevention. This paper systematically elaborates on the core pathways, key models, and implementation frameworks of mine safety management driven by big data. The study proposes a technical framework centered on “data perception-model-driven-intelligent decision-making”, and constructs an application path that focuses on a closed-loop system integrating multi-source IoT perception, dynamic risk modeling, and human-machine collaborative decision-making. Building on this, it focuses on three core application modes: the “digital twin + plan deduction” mode, which achieves advanced risk pre-control and emergency plan optimization through virtual-real integration and dynamic simulation; the “predictive maintenance + equipment health management” mode, which enables early fault warning and full lifecycle health management based on equipment operation data and predictive models; and the “behavioral profiling + intelligent supervision” mode, which utilizes positioning and AI technologies to quantify personnel safety characteristics, enabling precise supervision and risk warning. Furthermore, it proposes a key implementation framework comprising five layers: multi-source perception, data governance, intelligent modeling, decision application, and iterative optimization, clarifying the entire process from data collection and governance analysis to intelligent decision-making and continuous optimization. This research provides systematic theoretical support and a practical paradigm for the intelligent transformation of mine safety management, holding significant importance for enhancing the intrinsic safety level of mines.