Abstract: Driven by the urgent demand for intelligent development in coal mining, snake-like robots were considered as key equipment for addressing operational challenges in complex underground environments due to their bio-inspired mechanical structures and flexible transmission characteristics. As a novel type of bionic robot, snake-like robots were developed to overcome the motion limitations of traditional rigid manipulators, and were characterized by high flexibility and adaptability to explosion-proof requirements in coal mines, demonstrating significant application potential in underground inspection, equipment maintenance, and emergency rescue scenarios. The theoretical foundations and key technologies of snake-like robots were systematically reviewed. The development history and current research status were summarized, and the structural classifications and characteristics were analyzed to clarify the advantages and applicable boundaries of different types of snake-like robots. Based on this, the characteristics of discrete, serpentine, and continuum configurations were comparatively analyzed. At the theoretical level, kinematic and dynamic modeling methods, as well as control strategies for hyper-redundant systems, were systematically reviewed, with emphasis placed on the challenges posed by nonlinearity, dynamic uncertainty, and environment interaction, as well as the requirements for robust control and degraded operation mechanisms. At the level of key technologies, research progress in model-based and data-driven control methods, multi-modal perception fusion, and localization and navigation under communication-constrained conditions was summarized, and their engineering applicability in complex coal mine environments was analyzed. However, it was found that snake-like robots still faced several challenges in industrial applications in coal mines, including the trade-off between explosion-proof design and lightweight structure, reliability constraints, stable fusion of multi-source sensing information under complex conditions, and insufficient capability in high-precision localization and autonomous decision-making under communication limitations. To address these challenges, key development directions were proposed from the perspective of coal mine application requirements, including multi-functional integration and collaboration, intelligent control and autonomous decision-making, and efficient energy and actuation management. These findings were expected to provide systematic references for engineering applications and future research of snake-like robots in confined underground spaces. It was suggested that future studies should further focus on typical coal mine operational scenarios, and continuous efforts should be made in explosion-proof lightweight design, reliable perception under degraded conditions, and risk-constrained autonomous control, so as to provide stable and efficient core equipment support for intelligent coal mining and promote the advancement of the coal industry toward intelligent development.