The paper considers the problems that are associated with the design of building fault detection and control systems. Direct and indirect approacher to the design of controllers and fault detectors are described .A number of proposed schemes are compared before the key issues are identified . The advantages of a neurouzzy model-bases approach are then explained . Work on current research projects ,　 which are using this approach to develop a robust fault diagnosis scheme for air -handlers , an air temperature sensor for heating ,ventilation of air -conditioning applications , a new type of thermal comfort control scheme , and a fault-tolerant supervisory control scheme for air-conditioning systems ,is also reviewed.