This paper concerns nonlinear temperature control of a batch polymerization reactor where suspension polymerization of methyl methacrylate (MMA) takes place. For this purpose, four control algorithms, namely, a fix proportional-integral (PI) controller, an adaptive proportional-integral-derivative (PID) controller and two globally linearizing control (GLC) schemes, one for known kinetic model (GLC-I) and the other for unknown kinetic model (GLC-II), are selected. The performances of these controllers are compared through simulation and real-time studies in the presence of different levels of parameter uncertainty. The results indicate that GLC-I and GLC-II have better performances than fix PI and adaptive PID, especially in case of strong gel effect. The worst performance belongs to adaptive PID because of rapid model changes in gel effect region. GLC-II has a simpler structure than GLC-I and can be used without requiring the kinetic model. In implementation of GLC-I the closed loop observer should be used because of model uncertainties.