The aromatics complex is essential in the production of p-xylene and contains several distillation towers and reactors, making this process highly energy intensive. To save energy for the aromatics complex, a novel framework for optimizing the process and heat exchanger network simultaneously is presented. Two models are included in this framework, namely, the process model, which is established through artificial neural networks, and the heat exchanger network model, which includes streams with phase changes. In addition, a disjunctive formulation is proposed to diagnose the process streams with or without latent heat according to the operating conditions. The performance of the proposed approach is illustrated with two case studies. The results demonstrate the significant potential in energy savings and considerable increase in the profit of the whole system.