This article proposes the use of Output Error Model Predictive Control (OE-MPC) as  a standard MPC technique for temperature control applications. Least squares of an output error criterion embedded with MPC was implemented in the basis of the control law design.  The OE model was exploited to minimize the plant-model mismatch, and therefore the controller accuracy is significantly improved.  The paper compares the performance of the suggested controller with the performance of MPC based ARX (Autoregressive with exogenous input) model, and Generic Model Control (GMC) in temperature control of an exothermic batch process simulation. The work identifies the features of the considered algorithms in a comparative approach.