Artificial intelligence and knowledge-based systems have been applied to control systems to address the increasingly complex problems but controversy and confusions have remained unsettled. This paper takes a problem-dependent view, which is advantageous over a model-dependent one, and presents a systematic study on the various works under the term intelligent control, and particularly their profound implications. For a consensus of discussions, first of all, concepts of systems and control are generalized and rationalized to be applicable to unconventional types of control systems. Then, three distinctive types of problems are characterized, i.e., novel schemes of continuous-variable real-time (CVRT) control and its parametric adaptation, non-CVRT control, and hierarchical control architectures of large complex systems.