In order to design more operable or even optimal chemical processes, reliability theory is proposed in this paper to support planning of operating procedures. A three-tier reasoning framework is constructed to describe the two-step quantitative planning strategy. The first step is determination of optimal operating path, and the second step is generation of optimal primitive operations. Based on the deficiency of the NLP method in obtaining an optimal operating path, this paper presents the knowledge-model (K-M) method in which expert knowledge base is combined with process dynamic model to solve this problem. Finally, the methodology is illustrated by an example of raising temperature procedures for a subsystem of charge gas dehydrator regeneration in an ethylene plant.