Presented is a radiometric sensor and associated electromagnetic propagation models, developed to facilitate non-invasive core body temperature extraction. The system has been designed as a close-proximity sensor to detect thermal emissions radiated from deep-seated tissue 1 cm – 3 cm inside the human body. The sensor is intended for close proximity health monitoring applications, with potential implications for deployment into the improved astronaut liquid cooling garment (LCG). The sensor is developed for high accuracy and resolution. Therefore, certain design issues that distort the close proximity measurement have been identified and resolved. An integrated cavity-backed slot antenna (CBSA) is designed to account for antenna performance degradation, which occurs in the near field of the human body. A mathematical Non-Contact Model (NCM) is subsequently used to correlate the observed brightness temperature to the subsurface temperature, while accounting for artifacts induced by the sensor’s remote positioning from the specimen. In addition a tissue propagation model (TPM) is derived to model incoherent propagation of thermal emissions through the human body, and accounts for dielectric mismatch and losses throughout the intervening tissue layers. The measurement test bed is comprised of layered phantoms configured to mimic the electromagnetic characteristics of a human stomach volume; hence defines the human core model (HCM). A drop in core body temperature is simulated via the HCM, as the sensor monitors the brightness temperature at an offset distance of approximately 7 mm. The data is processes through the NCM and TPM; yielding percent error values < 3%. This study demonstrates that radiometric sensors are indeed capable of subsurface tissue monitoring from the near field of the body. However, the following components are vital to achieving an accurate measurement, and are addressed in this work: 1) the antenna must be designed for optimum functionality in close proximity to biological media; 2) a multilayer phantom model is needed to accurately emulate the point of clinical diagnosis across the tissue depth; 3) certain parameters of the non-contact measurement must be known to a high degree of accuracy; and 4) a tissue propagation model is necessary to account for electromagnetic propagation effects through the stratified tissue.