Geophysical field surveys can be used for mapping soil properties such as clay and water content. For this the choice of adequate geophysical pedo-transfer functions (GPTF) is a still debated question because of ambiguities of influencing factors. To investigate the applicability of different GPTFs to typical central European soils, we conducted laboratory measurements on 23 samples. The results are datasets of electrical resistivity and dielectric permittivity as a function of water content and partly temperature. We used these data to validate and improve existing GPTFs. For the relationship between permittivity and water content, we applied published empirical regressions, volumetric mixing, and effective media models. All models show low RMSE (0.3−6.8l%, v/v), but the curvatures of the data points are not reproducible by the empirical models. The volumetric mixing and the effective media models can fit the trend of the data points for sandy soils but not for clayey soils. To improve these models, we propose a modification to the Hanai–Bruggeman-type effective media approach. It uses the permittivity of almost dry soil as a background and embeds stepwise increasing amounts of water. It turned out that our approach yields satisfying fits with respect to both data trends and RMSE (0.3–1.5l% v/v). Based on this approach we define a joint model for permittivities and resistivities, which yield RMSE of only 0.7 to 1.4% (v/v)—better than most published models. Our validations show that the temperature effects on electrical resistivity can be reliably corrected using existing empirical relationships.