The unsaturated soil hydraulic conductivity K(θ) is quite important and required for many hydrological processes and soil-water flux modeling, and agricultural applications, including land-use evaluation, estimate of drainage, chemical leaching, and others. Measurements of  K(θ) are difficult, laborious, time consuming, and highly variable. These problems necessitating large number of sampling and measurements to achieve reasonable values of  K(θ) or to characterize an area of land. This will be costly and economically not feasible. Rather than measuring the K(θ) directly, different predictive models were used to estimate K(θ). In the recent years, intensive uses of pedotransfer functions as an approach to translate simple characteristics of soil surveys into more complicated  parameters, i.e. K(θ). We approach local best-fit of data models that accurately predict K(θ). Such models are rarely found and quite needed for the Egyptian soils to be used in several applications and models. There are numerous models of estimating K(θ). We evaluated K(θ) for three major Egyptian soils: alluvial-lacustrine, calcareous, and sandy using different predictive models. The proposed models:Best-fit of data (BFD), Campbell , 1974 (CAM),  Gilham et al., 1976 (GIL), Mualem-van Genuchten, 1980 (MVG), Rawls and Brakensiek, 1989 (R&B), and  Saxton et al., 1986 (SAX). Data showed that the BFD model best described K(θ) for the given soils. Other predictive models showed variations in their errors and precisions with different types of soils. The relative magnitude of the highest possible mean error (ME) were 2.6, 28.6, 7.8, 89.6, 93.5, and 100٪  for alluvial-lacustrine soil; 3.0, 52.7, 11.8, 78.8, 82.7, and 100٪ for calcareous soil; 6.0, 79.3, 84.7, 92.4, 100, and 66.0٪ for sandy soil; and 5.1, 68.7, 55.6, 93.4, 100, and 86.4٪  for all over types of the studied soils with the BFD, GIL, MVG, CAM, R&B, and  SAX models, respectively.