This study investigates the impact of moisture redistribution during the measurement of the thermal dry-out curve (i.e., soil thermal resistivity, ρ T , versus moisture content, θ). Thermal dry-out curves are used in energy geotechnics (e.g., geothermal heating and cooling, collector systems at wind energy sites) to quantify the capacity of a soil to store or dissipate heat. Although there is not a standard method for measurement of the ρ T -θ relationship, curves are typically generated by measuring ρ T of remolded or undisturbed specimens at varying water contents, typically with incremental measurements as the specimen is dried. However, drying within a specimen may not be uniform, even in a controlled, low-temperature, drying environment. In this study, a thermal properties probe was used to measure ρ T of remolded cylindrical specimens as the soil was dried in a low-temperature oven. Three horizontal sensor locations (top, middle, and bottom) were used to measure ρ T across the specimen. Water content at the sensor locations was measured and compared to the corresponding ρ T measurement. Thermal resistivity tests were performed at various drying times for three soils: Poorly Graded Sand (SP), Silty Sand (SM), and Sand with Silt (SP-SM). Tests on SM remolded at 95% maximum dry density revealed θ up to nine times higher in the bottom sensor location than in the top sensor location. The test results indicate a need for modification of the standard thermal resistivity testing procedure such that moisture migration within a specimen during drying does not impact the testing accuracy.