High frequency dielectric measurements of soil water content can exhibit temperature sensitivities inconsistent with that expected for bulk water. These sensitivities are significant in fine-textured soils and controlled by the interaction among the temperature dependencies of the static permittivity of bulk water, the volumetric fraction of bound water, and the bulk DC electrical conductivity (EC) of soils. We developed and evaluated a physically-based algorithm to calibrate TDR water content measurements in fine-textured soils. Our algorithm estimates the real and imaginary components of the complex dielectric permittivity using four component dielectric mixing models with bulk EC evaluated separately using TDR-measured signal attenuation. Temperature effects on the relaxation time for bulk and bound water were also included. Lastly, the decline in effective frequency during pulse propagation was estimated as a function of the calculated attenuation rate. Three replicates of the Ap and Bt horizons of a Pullman clay loam (fine, mixed, superactive, thermic Torrertic Paleustolls) and Richfield silt loam (fine, smectitic, mesic Aridic Argiustoll) were packed in 0.2 m diameter PVC columns and adjusted to water contents ranging from air-dry to near saturation. Travel time and attenuation of the signal were measured using a Tektronix 1502C cable tester at 8, 23, and 40 ¢ªC and using three coaxial cable setups with 0.2-m trifilar probes. Reference air, water, and short-circuit probe measurements were completed for each cable setup. Two and three-parameter nonlinear fits of apparent permittivities closely approximated measured permittivities for each soil horizon, temperature, and cable setup. The calibrations were valid up to a maximum estimated loss tangent of 1.5 for the Pullman Bt horizon at saturation with a minimum effective frequency of 80 MHz and a maximum bulk EC of 0.15 S/m.