A methodology is proposed to target the retrofit of utility system and heat exchanger network (HEN) of large-scale chemical plants. The methodology is based on the Mixed Integer Linear Programming (MILP) formulation for the optimal site utility integration and combined heat and power production. In order to identify the best degree of site integration in the retrofit problem, the MILP model is extended by introducing the entire associated heat transfer interfaces of each energy requirement. In addition, to find a region of good solutions, a multi-objective optimization is performed to generate the optimal Pareto-frontier by minimizing both operating and investment cost. Formulating the complete retrofit problem as an optimization problem allows identifying prior to the HEN design step, the heat transfer units which have to be modified to reach the energy target and thus it prevents the iteration between targeting levels. Moreover, it considerably reduces the size of HEN design problem. The methodology is applied on a real large-scale plant with more than 170 process streams and its practice is highlighted by showing a solution with 50 % lower investment cost and number of streams for HEN model than original MILP while it has only 3 % higher operating cost.