Adequate optimization formulation and algorithm choices are essential for successful optimal design of a multidisciplinary system. Details of such process remain with the realm of tacit experimental knowledge and skills of engineers. The future goal of our research is to build a descriptive management method that facilitates explicit description of experimental knowledge of engineers. This paper aims at clarifying an acquisition process of experimental knowledge in optimal design of a multidisciplinary system. An experimental design optimization of a desiccant air system is performed to capture the process. Some typical process patterns are formalized based on the analysis of the captured process. This paper also discusses a vision for the descriptive management.