Abstract: The design of energy efficient water utilization systems becomes a recent research direction. However, even for utility targeting, the current models involve many variables and constraints and it takes much time to solve the problem. Furthermore, the resulting utility target relies on the assumption ofΔTmin=0℃, for treating both direct and indirect heat transfer in one transshipment model. This assumption may lead to ΔTmin violations in the following HEN construction. A new kind of NLP model for utility targeting underΔTmin=0℃ is presented which can reduce the number of variables greatly. Moreover, an improved design approach considering the specifiedΔTmin is generated. This approach, which treats the direct and indirect heat transfer separately, provides the information to construct the water utilization system as well as the corresponding heat exchanger network without anyΔTmin violation.