ABSTRACT This paper discusses a methodology for accurately estimating permeability from well log devices, with commonly available core and log data. Core permeability and porosity data are used to identify three or four different hydraulic zones (HZ's) with unique pore structures. Logging tools are selected that show different responses for each hydraulic zone. Multidimensional databases are constructed that relate three or four logging values to core permeability to provide relationships for permeability prediction from logs where core is not available. Predictions are more accurate than with other methods because permeability transforms are provided for each pore type and because pore throat size is considered. Well log permeability estimations derived by this method are used to map permeabilities for improved reservoir simulation models, to improve well test permeability estimation, and to optimize well completions. This technique can be applied in either vertical or horizontal wells.