The refining and petrochemical industries generally own process plants and utility systems. Process plants are configured to finish the transformation and separation of materials, and utility systems supply the energy requirements for the process plants. Therefore, integrating two of them is more favorable than optimizing them individually. A coupling mixed integer nonlinear programming model is presented in this work to integrate process plants and utility systems; the objective is to minimize the energy costs to meet the requirements of the process operations and to maintain a steam balance in the total site. The mathematical model includes three parts: the heat integration of the process plants, the optimization of the utility system, and the coupling equations for the site-scale steam integration. The heat integration of the process plant is formulated on the basis of pinch analysis involving heat loads of the process heaters and steam generation and requirements. An optimization of the utility system is also proposed to provide the relationship between steam balance, power generation, and fuel requirements. Coupling equations are used to balance the steam streams in each level between the process plants and utility systems. Two real industrial examples are also investigated to demonstrate the performance of the presented mathematical model. The solution results indicate not only a more profitable integration scheme but also increases in energy utilization efficiencies and the operational capacities of the utility systems.