Abstract: Accurate perception of precursor information for coal spontaneous combustion in goaf is a critical prerequisite for achieving intelligent disaster monitoring and advanced early warning. However, the significant medium absorption and multipath attenuation effects in the porous media environment of goaf severely limit the wireless transmission distance of disaster monitoring information, restricting the practical application of existing wireless intelligent sensing technologies. Therefore, addressing the challenge of stable long-distance wireless signal transmission in goaf porous media environments is fundamental to enabling intelligent monitoring and advanced early warning of coal spontaneous combustion in goaf. This study employs frequency-domain finite element numerical simulation and establishes electromagnetic equivalent models for both goaf and soil porous media based on effective medium theory. The effects of antenna polarization and operating frequency on wireless signal attenuation in these two types of media were simulated and investigated. Concurrently, through soil porous media experiments, the wireless signal propagation characteristics under various antenna lobe widths and other radiation parameters were systematically tested. Based on the classical log-normal shadowing loss theory, path loss under different antenna parameter combinations was uniformly characterized and compared. The results show that the established electromagnetic equivalent models of goaf and soil porous media can effectively simulate and experimentally analyze the influence of antenna radiation characteristics on wireless signal attenuation in goaf environments. By incorporating correction terms for polarization mismatch, frequency-dependent medium absorption, and lobe scattering, a path loss model capable of describing the attenuation patterns of wireless signal propagation in complex porous media was developed. Signal intensity exhibits a logarithmic decay trend with propagation distance, with linearly polarized antennas demonstrating greater robustness in porous media, and their average polarization mismatch loss is approximately 1.4 dB, significantly lower than that of circularly polarized antennas. Under the same propagation distance, signal attenuation increases with frequency. The 170-230 MHz frequency band exhibits the lowest attenuation and optimal propagation performance, while the 433 MHz band balances attenuation characteristics and antenna size, and the 868 MHz band suffers significant attenuation and is therefore unsuitable for deep goaf communication. In soil porous media, antenna coverage capability correlates positively with lobe width; narrow-lobe antennas have limited offset tolerance, whereas wide-lobe antennas are more conducive to achieving spatial coverage and multipath transmission. By constructing a wireless signal attenuation model and path loss characterization method for goaf porous media, this study quantitatively investigates the effects of polarization mismatch, medium absorption, and lobe scattering on wireless signal transmission in goaf porous media. The findings provide theoretical foundations and technical support for wireless communication link design and node deployment in intelligent perception and advanced early warning systems for coal spontaneous combustion in goaf.