Responding to the technological progresses and the emerging trends of chemical industry, the mode of intelligent chemical industry based on cyber-physical system (CPS) was discussed. A structure of CPS for intelligent chemical industry was proposed, followed by the key techniques, including big data techniques and rigorous online modeling. As an example, the intelligent rectification column was elaborated. The structure of the three-agented column was proposed, to improve the optimal control level for the whole process system and promote on-line optimizing control techniques with robustness, in which CPS was the basic platform delivering the functions of surveying, communication, data storage, and computation. The optimal control models of distillation for the whole system were deduced as well; the assessment algorithms of equipment operating status, anomaly recognition, and multi-device collaboration based on the agent-oriented method were also illustrated. The operation results from the propylene rectification column showed that the intelligent rectification column proposed in the study was feasible, and met the operational requirements in terms of feedback time, stability, robustness etc.