Lack of soil physical data, particularly soil water release data and hydraulic conductivity data, is recognised as one of the greatest limitations to the widespread application of simulation models, needed to address environmental issues. Because of the expense of generating new soil physical data pedotransfer functions may be used to predict soil physical data from existing information, notably soil morphology. Pedological horizon descriptions can then be used to estimate soil physical properties for many points in the landscape. The soils used in this study were derived from a systematic sampling of soil profiles for soil physical characteristics for 8 soils series within 2 drainage sequences on the post-glacial and glacial surfaces of the Canterbury Plains. Soil series in each sequence varied from shallow sandy loam, well-drained soils to deep clay loam, poorly drained soils. Each soil series was represented by 9 profiles. Three horizons in each soil profile were sampled for soil porosity values, particle size, and saturated- and near-saturated hydraulic conductivity. Pedological horizons were grouped into functional horizons on the basis of soil morphologic attributes expected to have closest relationships with soil physical properties (ped size, ped type, packing class, consistence and presence of argillic horizons). For topsoils, functional horizons based on ped size were found to have greatest predictive ability and provided separation between horizons for bulk density, macroporosity, clay content, wilting point, readily available water, and near-saturated hydraulic conductivity. For subsoils, horizons with clay content >35% had distinct relationships with soil physical properties and needed to be separated from other subsoil horizons. For the remaining horizons, separations in soil water release characteristics and some hydraulic conductivity data were obtained by functional horizons based on packing class and the presence of argillic horizons. Adding ped size to the functional horizon definition provided further separation of horizons for hydraulic conductivity. This study demonstrates that a range of pedological horizons, derived from a wide range of soil types, can be grouped into 4 functional topsoil horizons and 3–5 functional subsoil horizons on the basis of simple morphological attributes or merged pedological horizons. The functional horizons, thus created, enable statistical distributions of soil water release and hydraulic conductivity data to be predicted for map units.