Two water network optimization methods are presented. Both can be utilised to establish, without detailed design, whitewater network configurations that would minimize fresh water requirements and waste-water rejects while respecting critical contamination constraints. The essentially graphic pinch-type analysis is partially automated; it is only applicable to one contaminant system. Linear programming appears to be a more powerful tool because it can be applied simultaneously to several contaminants to be removed and fine fibers to be retained in the system while fresh water consumption is minimized, It is fully automated and particularly well adapted to large size problems. Both methods have been applied to industrial case studies with interesting results. In the case of the water network of the alkaline zone of a de-inking plant no fresh water consumption reduction was obtained, but it was shown that network rearrangement could eliminate the need for a filtration step. In the case of the complete whitewater network of an integrated newsprint mill, it was shown that it is possible to reduce the fresh water consumption by two thirds.