This paper centers on a new knowledge-based system (KBS) for diagnosing the surface condensation caused by numerous conditions such as the thermal performance of buildings, building facilities, climatic conditions and consideration of the occupants＇ comfort, and suggests appropriate solutions for the moisture problems regarding houses. The KBS consists of four primary modules, i.e., user interface, system controller, function program and database. Each module includes several subcomponents, for instance, the function program module has an inference engine, file controller, procedure explanation, data collector and so on. The simulation program for predicting room temperature, surface temperature, humidity and ventilation rate is integrated into the function program module. Therefore the duration and position of condensation in a room can be estimated using this program. After diagnosing the moisture problems using the rule-base knowledge and calculation modules, the problem-solving strategies are evaluated and presented quantitatively during the last stage. Moreover, a consultation example of a detached house built in a cold climatic region in Japan is shown and the validity of this system is verified.