This work deals with the temperature control in a fluidized bed reactor (FBR). The strategy is an input-output linearizing feedback control scheme which involves an uncertainties dynamic estimator. The developed dynamic compensator allows accurate temperature regulation in spite of uncertainties associated with modelling errors. As the uncertainty estimation converges to the actual uncertainties, the controller achieves asymptotic input-output linearization. The control strategy for the FBR based upon this scheme resulted in good performance on temperature set-point tracking.