Water is widely used as a solvent in the mining industry and is employed in hydrometallurgical processes and for mineral concentration. Because of the global increase in metal production, the demand for water, including fresh water, is expected to increase continually. In arid and semi-arid regions such as northern Chile, the scarcity of fresh water has led to increased dependence on other sources such as sea water and triggered efforts towards optimization of the use of fresh water. In copper concentration plants, approximately 40-60% of the total amount of water lost is retained in slurries in the tailings. In this paper, we present a method for optimizing the design of dewatering systems that employ hydrocyclones and thickeners. Mathematical models were generated to determine the maximum water recovery rate and the corresponding system structure for given equipment sizes, and to determine the minimum cost of the equipment and the corresponding system structure for given water recovery rates. The models were based on mixed integer nonlinear programming. Several case studies were performed. The model predictions were consistent with the results of an experimental study of an actual dewatering system in a copper concentrator plant.