A new and simple method is presented for estimating soil water diffusivity. This method utilizes general similarity theory rather than the Boltzmann transformation to evaluate horizontal water infiltration-re-distribution processes. It uses the Brooks and Corey function of water diffusivity and requires only measuring the wetting front advance with time. General similarity diffusivities for five soils were compared with those obtained by the Boltzmann transformation and with a third method that used a fitting function to approximate the water distribution data in the Boltzmann transformation method. The comparisons showed that soil water diffusivities for the three methods were in good agreement for the intermediate range of water contents. However, at low water contents, the similarity water diffusivities differed from the other water diffusivities for the five soils. The new method has several advantages over the other methods. It allows the inlet boundary water content to vary in time and initial water content distribution to vary with distance, which is more general than constant water content. The new method also does not require soil water diffusivity to be zero at the initial water content. This represents an improvement over the earlier methods, which give a zero diffusivity at initial water content no matter how high this initial water content is.