Accurate estimation of water saturation is central in predicting capillary pressure and relative permeability, under special core analysis in the laboratory. We have explored the use of dielectric measurements at different frequencies to estimate water saturation. In addition to water saturation, dielectric measurements are sensitive to the distribution of water and oil in a porous system, reflected by the apparent cementation factor [Formula: see text], which describes the water phase tortuosity. We have performed an experimental study to benchmark water saturation from dielectric measurements on eight carbonate cores and estimated their cementation exponent [Formula: see text] and saturation exponent [Formula: see text] in Archie’s equation from dielectric data. All cores went through a series of drainage/imbibition steps, creating varying saturations of brine/fluorocarbon. Fluorocarbon was chosen because it is invisible to proton nuclear magnetic resonance (NMR). Therefore,NMR porosity represents only the water-filled porosity and can be used to benchmark dielectric water-filled porosity. Three dielectric models were used for the comparison of the dielectric water-filled porosity with the one from NMR, i.e., the complex refractive index model (CRIM), bimodal model, and Stroud-Milton-De (SMD) model, and very good agreement of 1.5 porosity units on average is found. Despite its simplicity, CRIM predicted well the water-filled porosity in this experiment. However, it cannot provide information about the texture, which is captured by bimodal and SMD models. We also estimated [Formula: see text] and [Formula: see text] based on [Formula: see text] found from bimodal and SMD models, and good agreement with [Formula: see text] from resistivity data was shown. This is the first time to our knowledge that such a rich set of dielectric and NMR measurements was acquired at different saturation stages in a surface laboratory. This study is useful in benchmarking the water saturation from dielectrics, comparing different dielectric models, and demonstrating feasibility of estimating textural parameters.