Conventional architectural design has a lot to do with the intuition and experience of designers. And residences are not always suit to its residents and surrounding environment. If we can extract residents' preferences and demands about comfort of each resident from histories of past life and reflect these information in next design, it's possible to make living space more comfortable. This thesis proposes genetic and evolutional system for architectural design information, which is applied evolutionary adaption. Specifically, I applied genetic mechanism which base sequence of DNA plays a role, epigenetic mechanism which chemical modification plays a role and evolutionary mechanism with natural selection. Proposed system firstly accumulates discomfort of residents, shortcoming of living space and usage of equipment as "comfort stress", "safety stress" and "energy saving stress", and modifies performance value of related performance items of building depending on the stress accumulation. Then this system processes selection according to the characteristics of the site for candidates of next generation of architectural design information which are generated via crossing and mutation. The data-set selected in this way is regarded as the performance value of next architectural design, and system suggests architectural specification to the residents.