Industrial water networks are designed in the first part by a multiobjective optimization strategy, where fresh water, regenerated water flow rates as well as the number of network connections (integer variables) are minimized. The problem is formulated as a Mixed-Integer Linear Programming problem (MILP) and solved by the ε-constraint method. The linearization of the problem is based on the necessary conditions of optimality defined by Savelski and Bagajewicz (2000). The approach is validated on a published example involving only one contaminant. In the second part the MILP strategy is implemented for designing an Eco-Industrial Park (EIP) involving three companies. Three scenarios are considered: EIP without regeneration unit, EIP where each company owns its regeneration unit and EIP where the three companies share regeneration unit(s). Three possible regeneration units can be chosen, and the MILP is solved under two kinds of conditions: limited or unlimited number of connections, same or different gains for each company. All these cases are compared according to the global equivalent cost expressed in fresh water and taking also into account the network complexity through the number of connections. The best EIP solution for the three companies can be determined.