The liberation characteristics of any ore are intimately related to the mineralogical texture. In this paper, leaching experimental results were used with the liberation model of King to predict the liberation of gold from complex ores. Diagnostic leaching results were used with the liberation model, as the results from a diagnostic leaching procedure explain the deportment of gold in the various minerals in a complex ore. The free or leachable gold in an ore obtained from such an experiment would be similar to the gold liberated by crushing and/or milling. The leachable gold in each size fraction of an ore can be determined by diagnostic leaching, before and after milling. Thus, it would be very useful if the leachable (liberated) gold in a specific unmilled ore can be used to predict leachable (liberated) gold in the milled ore of that specific ore. If this can be achieved, then the leachable gold in each size fraction of the unmilled and milled ore can be predicted by the model. An explicit solution to King＇s liberation model was modified in order to predict the leachable gold in different size fractions on the basis of data gathered for the unmilled ore. In this model, mu(m) is the mean linear intercept length for gold in the ore, and the term A.exp(1/D) indicates the fraction of leachable gold which cannot be attributed directly to liberation: L(g)(D) = 2mu(m)/D2[mu(m) -(mu(m) + D)exp(-D/mu(m))] + [A.exp(1/D)] It was found that this model gives reasonable predictions of the leachable gold for milled ore when the leachable gold in the unmilled ore is known.