Abstarct: Abstract In this study, a Robust Model Predictive Control (RMPC) scheme is designed and implemented for a non isothermal Continuous Stirred Tank Reactor (CSTR). Because of its strong nonlinearity, multiplicity, and sensitivity to temperature and kinetic parameters, the non isothermal CSTR represents one of the most challenging benchmark processes in process control. First, the nonlinear mass and energy balance equations of the CSTR were derived and linearized around a stable operating point to obtain a state space model. A conventional Model Predictive Controller (MPC) was then designed as a baxseline. Subsequently, to cope with simultaneous parametric uncertainties in F, k0, E,F, k_0, E,F, k0, E, and TfT_fTf, a min–max RMPC approach with polytopic uncertainty descxription was utilized to guarantee robust stability and constraint satisfaction under worst case scenarios. Five uncertainty cases were simulated, and both MPC and RMPC controllers were evaluated under identical conditions. The RMPC achieved a 35 % reduction in the average cost function JavgJ_{avg}Javg, improved the settling time from 33.8 s to 28.3 s, and increased process stability up to 95.6 %. The tracking error e(t)e(t)e(t) was consistently smaller, providing smoother and more reliable reactor temperature responses across all cases. Overall results confirm that the RMPC strategy offers a reliable and efficient control frxamework for nonlinear chemical reactors under parametric uncertainties and can be extended to complex industrial processes such as polymerization reactors.