The NECTAR-project (Neural and Expert ConTrol of AiRcraft) aims to investigate the application of artificial intelligence techniques to flight control. This paper focuses on part of this research, namely the application of a real-time expert system in automatic flight control. The expert system has been developed to control the landing procedure of a twin engined business jet under a variety of conditions. This application requires fast and predictable response times from the expert system. To achieve this, a state-based approach has been adopted and a system of knowledge-base compilation has been used. In the current system, low-level control of the flights surfaces is performed by convetinal controllers, however one of the aims of the NECTAR project is to investigate the use of neural networks-based controllers in this role. Results are presented in this paper that were obtained from simulated flight tests.