This paper presents mass balance calculations using genetic algorithms for copper smelting in an Outokumpu fl ash furnace. Based on the elemental composition of the copper concentrates being fed to the reactor, the mineralogical composition of the concentrate mixture is adjusted by means of genetic algorithms. The macroscopic mass balance equations for the species entering and leaving the furnace are solved and the compositions and fl ow rates of matte, slag, and the off-gas stream are computed. Good agreement between the predicted and plant data was obtained in terms of matte and slag fl ow rates, matte grade, and copper, iron, magnetite, and silica contents in the slag. Predictions are more suitable and faster to obtain with this method than a conventional method in which the mineralogical composition of the feed is not adjusted. Future applications of the formulation are discussed.