Abstract: Driven by the urgent demand for intelligent coal mining, snake-like robots, benefiting from bio-inspired serpentine structures and flexible transmission mechanisms, have emerged as a key solution for operations in complex underground environments. As a new type of biomimetic robot, snake-like robots overcome the motion limitations of traditional rigid manipulators and combine high maneuverability with adaptability to explosion-proof requirements in coal mines, demonstrating significant potential in underground inspection, equipment maintenance, and emergency rescue scenarios. This paper systematically reviews the theoretical foundations and key technologies of snake-like robots, tracing their development history and research progress. Starting from structural classification and characteristics, the advantages and applicability boundaries of different types of snake-like robots are summarized, and their application explorations in the coal mining field are comprehensively analyzed. On this basis, the characteristics and applicable scopes of discrete, articulated (meandering), and continuum configurations are discussed. At the theoretical level, kinematic and dynamic modeling methods as well as control strategies for highly redundant systems are reviewed, highlighting the need for robust control and degraded-operation mechanisms under nonlinear dynamics, uncertainty, and strong environment interaction. At the key-technology level, advances in model-driven and data-driven control methods, multimodal perception fusion, and localization and navigation under communication-constrained conditions are summarized, with an emphasis on their engineering applicability in complex coal mine environments. Nevertheless, the industrial application of snake-like robots in coal mines still faces major challenges, including engineering trade-offs between explosion-proof design, lightweight structures, and reliability; stable fusion of multi-source sensing information under harsh conditions; and insufficient high-precision localization and autonomous decision-making capabilities under communication constraints. In response to these challenges, this paper identifies key development directions from the perspective of coal mine application requirements, including multifunctional integration and system coordination, intelligent control and autonomous decision-making, and efficient energy and actuation management, providing systematic references for engineering applications and future research of snake-like robots in underground confined spaces. Future research should further focus on typical coal mine operational scenarios, achieving continuous breakthroughs in explosion-proof lightweight structures, trustworthy perception under degraded conditions, and risk-constrained autonomous control, to provide stable and efficient core equipment for intelligent coal mining and to promote the advancement of coal mine intelligence.