Estimation of permeability in a reservoir is necessary for simulation of production history. In mature fields, cores are limited so estimation of permeability is usually done from permeability-porosity correlations developed from cored wells. In this paper, a methodology is presented to predict permeability from well logs. The method groups the well logs into electrofacies, which are recognized to correspond with core lithofacies. The permeability-porosity correlations in the core data are developed using petrophysical classifiers - rock fabric number, rfn for dolomite and flow zone indicator, FZI for sandstone. The rfn and FZI are calculated for the cored wells and calibrated to the well log measurements within each electrofacies. The petrophysical classifiers extract the pore connectivity information from the well logs within each electrofacies, representing similar lithology. The permeability is then calculated from the porosity, using the petrophysical classifier values. Application of the methodology is demonstrated for dolomite and sandstone intervals in the Ogallah Field, Kansas. Permeabilities obtained from this approach closely matched core measurements.