The present paper describes the use of an intelligent Flame Monitoring System on regenerative steel reheating burners based on direct measurement and analysis of the flame radiation signals. A series of experiments were conducted on a 500 kW furnace fitted with two burners firing in a regenerative manner. The experiments covered a wide range of burner operating conditions including variations in the burner firing-rate and excess air levels. Gas supply to one of the burners was manually reduced in order to simulate burner imbalance. The flame radiation signals were acquired using a fibre-optic based optical instrument incorporating broad ultraviolet, visible and infra-red photodiodes. The correlation between the dynamic flame signals with respect to the excess air level and nitrogen oxides emissions were made using neural network models following off-line analysis of the acquired signals using different signal processing methods, to yield a set of flame features. The present work indicates that the measurement of flame radiation characteristics, coupled with advanced data modelling techniques such as neural network, provides a promising means of monitoring and optimising burner performance.