Abstract: Intelligent control is aimed to solve the task-oriented (goal seeking) control problems in complex environments which are unstructural, uncertain and have strong interactions with the plants. Traditional AI approach with emphasis on symbolic processing and logic reasoning has been confronted with serious difficulties in action oriented tasks, however, studying of new techniques for modelling and recognition of complex environments by mathematically precise ways and development of design techniques for imprecisely defined environments and tasks based on ideas in the newly emerging schools of AI may be integrated to provide new schemes and architectures for intelligent controls. In this paper, the authors anticipate the possibility that significant breakthrough would occur when the research following the top-down lines intersects with those following the bottom-up lines.