A new methodology is herein suggested to target the retrofit of utility system and heat exchanger network (HEN) of large-scale chemical plants for energy eciency improvement purposes. The methodology is based on the optimal site energy conversion unit integration and combined heat and power production, while still integrating the heat recovery identified by the heat cascade. A Mixed Integer Linear Programming (MILP) formulation is employed for this purpose. In order to identify the best degree of site integration in the retrofit problem, the entire associated heat transfer interfaces of each heating/cooling requirement is introduced to the MILP problem by defining each interface with one energy requirement representation which has a specific temperature level. A constraint is also added to the MILP formulation to specify that only one representation of each requirement is simultaneously allowed in the optimization problem. In order to find a region of good retrofit targeting solutions, a multi-objective optimization is performed to generate the optimal Pareto-frontier by minimizing both operating and investment cost. The application of the methodology will be illustrated on a real site scale process integration example including more than 170 hot and cold process streams.