Abstract: Monitoring mining-induced stress is a prerequisite for understanding stress redistribution and preventing dynamic disasters in deep underground engineering. However, deep mines and the coordinated extraction of coal-measure associated resources involve not only three-dimensional stress changes but also coupled seepage and thermal responses, whereas conventional monitoring systems usually measure isolated parameters at different locations and times. This review aims to clarify the performance boundaries of existing mining-induced stress monitoring methods, identify a technically suitable core for three-dimensional sensing, and propose a pathway toward integrated stress-pore-pressure-temperature monitoring. A structured comparative synthesis is conducted for five representative technologies: vibrating-wire stress meters, hydraulic stress meters, piezomagnetic sensors, fiber Bragg grating (FBG) sensors, and hollow inclusion strain cells. Their measurement mechanisms, monitoring dimensionality, installation and coupling requirements, temperature sensitivity, long-term stability, engineering applicability, and potential for multi-parameter integration are compared. For hollow inclusion strain cells, particular attention is paid to the stress-strain inversion principle, structural improvement, bonding-medium compatibility, strain transfer through the rock-cement grout-epoxy multilayer system, temperature compensation, and field performance. Existing correction strategies for multilayer strain transfer are further classified into experimental calibration, analytical correction, numerical inversion, and data-driven surrogate modeling. Evidence from representative applications in deep metal mines, coal mines, underground research laboratories, and hydraulic-fracturing projects is synthesized to evaluate practical feasibility and remaining uncertainties. The comparison shows that vibrating-wire and hydraulic devices are structurally simple and mature in field use, but their outputs mainly represent local pressure or one- to two-dimensional stress changes and usually require multiple sensors or boreholes for three-dimensional characterization. Piezomagnetic sensors provide convenient installation and possible self-diagnosis through redundant components, but their accuracy is sensitive to rock-mass integrity, prestressing, and installation coupling. FBG sensors are intrinsically safe, resistant to electromagnetic interference and corrosion, and suitable for remote, multiplexed, or distributed sensing; nevertheless, strain-temperature cross-sensitivity, packaging and coupling errors, and interrogation cost remain major constraints. Hollow inclusion strain cells have the most prominent advantage for this application because multi-directional borehole-wall strains measured in a single borehole can be used to invert both the magnitude and orientation of the three-dimensional mining-induced stress tensor. Their accuracy, however, is governed by installation quality, borehole disturbance, rock anisotropy, structural stiffness, bonding-medium properties, multilayer strain transfer, and temperature drift. Analytical and numerical studies indicate that stiffness mismatch and layer thickness can significantly alter strain transfer, while field applications demonstrate that hollow inclusion systems can continuously capture stress adjustment caused by excavation, mining, unloading, and hydraulic fracturing and can track changes in principal-stress magnitude and direction. On this basis, the review identifies poor spatiotemporal alignment, excessive borehole occupation, additional installation disturbance, incompatible communication protocols, and difficult data fusion as the principal limitations of separated single-parameter systems. An integrated architecture is therefore proposed in which a hollow inclusion strain cell acts as the three-dimensional stress-sensing core, while pore-pressure and temperature units are embedded at the same monitoring node. Co-located and synchronized acquisition would improve the consistency of multiphysics observations, allow temperature information to support both environmental interpretation and measurement compensation, and provide a stronger data basis for identifying coupled stress-seepage-temperature precursors. A three-stage development route is outlined: sensor selection and structural integration; development of multi-channel acquisition electronics, temperature compensation, data transmission, and host-computer software; and laboratory calibration followed by field validation and iterative optimization. Future work should focus on mechanically compatible and durable bonding media, physics-informed multilayer inversion, long-term drift control, intelligent sensor placement, multi-source data fusion, and digital-twin-assisted warning models. The main contribution of this review is to connect a mechanism-based comparison of stress-monitoring technologies with the multiphysics requirements of deep mining and coordinated resource extraction, thereby establishing hollow-inclusion-centered, integrated multi-parameter sensing as a practical direction for more reliable hazard identification and early warning.