This research evaluates the thermal properties of the proposed backfill materials (prepacked concrete) that function as the heat release medium for the underground power cable system (UPCS). First, a hot disk sensor was used to measure the thermal conductivity of small-scale backfill materials in the laboratory. The horizontal thermal response test (TRT) [1] was then performed with a 5 × 3 × 2 m full-scale sample for the in-situ evaluation. Afterward, the data from the TRT was utilized to couple with the infinite line source model (ILS) [2] to predict the thermal conductivity of the full-scale sample. Finally, the thermal conductivity of the newly developed proposed backfill materials was back-calculated using COMSOL Multiphysics. All these tests were also performed with natural sand (conventional backfill materials) to compare with the newly developed backfill material. Thermal conductivities calculated by numerical simulation and the ILS model were in very good agreement (difference < 3%), demonstrating that the proposed ILS model coupled with horizontal TRT data is appropriate for estimating the thermal conductivity of full-scale backfill material in situ. The three testing methods' results indicated that the prepacked concrete has a high thermal conductivity (> 2 W/(mK)), thus satisfying the heat release ability requirement for the UPCS. The laboratory test slightly underestimated the thermal conductivity (less than 8%) compared to the estimated values from the ILS model and numerical model involving prepacked concretes; however, natural sand showed a significant difference (1.365 W/(mK) and 1.8 W/(mK), 32 %) attributed to the influence of the water content change during the TRT. Therefore, it is recommended that the in situ testing conditions should be considered for the sand (or native soil) to avoid the overestimation or underestimation of their thermal conductivity.