An estimation of the effective thermal conductivity of the shallow subsurface zone in plains and basins where major cities have been formed is beneficial for the design of appropriate ground-source heat pump (GSHP) systems. Plains and basins consist of thick unconsolidated sediment formed during the Quaternary and Neocene periods in Japan. Therefore, sampling of the undisturbed geological core is required for the evaluation of the thermal conductivity of shallow subsurface zones; however, few of these samples have been collected in Japan. In addition, it is crucial to study the determining factors of thermal conductivity in undisturbed geological core samples. The aims of this study were to determine the thermal properties of unconsolidated and undisturbed subsurface samples for utilization in the design of GSHP systems, and to explore new factors characterizing the thermal conductivity of soil. We measured the thermal conductivity and volumetric heat capacity of five undisturbed boring cores of 50 m or 100 m deep sampled from the Tohoku and Hokuriku regions, the Tsugaru Plain, the Akita Plain, the Sendai Plain, the Yamagata basin, the Koriyama basin, and the Kanazawa Plain, using two type measurement sensors. Thermal conductivity, volumetric heat capacity, and volumetric water content were measured at intervals of a few centimeters in the Akita Plain core samples. For other core samples, thermal conductivity was measured at intervals of several dozen centimeters in both wet and dry conditions. In addition, porosity and pore size distribution were also measured at almost the same depth as the thermal conductivity measurement by using mercury intrusion porosimetry. The average thermal conductivity and volumetric heat capacity of all measured samples under wet conditions were approximately 1.48 W/(m·K) and 2.76 MJ/(m³K), respectively. The thermal properties tended to be related to the soil type. In all samples measured, the thermal conductivity in wet conditions showed weak negative correlation with the porosity. After selection of samples with many relatively large pores, the correlation between the porosity and the thermal conductivity of those samples was improved, and the correlation has well fitted by weighted harmonic mean model.