Conventional operations of wastewater treatment systems use the concepts of steady-state control, and often lead to unnecessary resource consumption for maintaining system functions. Real-time control was examined as a useful approach for improving the operation of wastewater treatment systems. This paper presents the application of real-time control to enhance the performance of nitrogen removal in a continuous-flow SBR system. A real-time control system combining on-line measurement of ORP and pH with Artificial Neural Network (ANN) model was proposed to carry out unsteady-state regulation of the hydraulic retention time of different operation phases. The result of this study shows that the performance of nitrogen removal was enhanced under real-time operation. Compared with fixed-time operation, the retention time of aerobic and anoxic phases can be reduced by approximately 45% and 15.5% in real-time operation respectively, also meaning that 45% aeration energy can be saved. The real-time operation also reveals a higher total nitrogen removal in a relative short retention time. Moreover, some dynamics and kinetics of nitrogen were investigated. These indicate the occurrence of nitrite-type nitrification under real-time operation. This nitrite-type nitrification results in the enhancement of denitrification performance with less carbon resource requirement and higher denitrification efficiency. (C) 1998 Published by Elsevier Science Ltd. All rights reserved.