This paper proposes a simple fractional order controller combined with a Smith predictor scheme for controlling the temperature of a steel slab reheating furnace. The dynamic model of the preheating zone of this process is obtained from an identification procedure applied in an industrial furnace. This identification procedure yields a second order plus time delay transfer function which undergoes large time delay changes. A fractional order integral controller combined with a Smith predictor is therefore designed. Simulated results compare the performances of the proposed fractional order controller with a standard PI controller, also combined with a Smith predictor, an LQR controller, and an H ∞ robust controller, in the case of the nominal process, and when the time delay varies. Four performance indexes have been used in this comparison: three related to the output performance (settling time, overshooting, and integral absolute error (IAE)), and a fourth one related to the control effort (TV). The analysis of these indexes shows that the simple fractional order controller provides lower values of the compared indexes when time delay becomes much higher than the nominal value. © 2013 Institute of Control, Robotics and Systems and The Korean Institute of Electrical Engineers and Springer-Verlag Berlin Heidelberg.