A chemical or petrochemical site is generally made up of several plants that are linked together through process streams. The linking process streams are often cooled down in their source plants, then transferred into storage tanks, and reheated in destination plants. This repeatedly cooling and heating results in low energy-use efficiency and more area installed in heat exchanger network. In this study, we introduce a heat exchanger network superstructure based on stage-wise model for heat integration using hot direct discharges/feeds between plants, and develop a new mixed-integer nonlinear optimization model to simultaneously design heat exchanger network. Unlike conventional HEN design, the model can simultaneously synthesize heat exchanger networks for multiple plants, and be able to address variable supply or target temperatures of process streams. The objective is to minimize total annual cost of heat exchanger networks in source and destination plants. Three examples are used to demonstrate the performance of the proposed model and solution approach. The computational results indicate that the simultaneous design of heat exchanger network for heat integration using hot direct discharges/feeds between plants achieves a significant decrease in total annual cost when compared to the separate design of heat exchanger networks for source and destination plants. (C) 2016 Elsevier Ltd. All rights reserved.