This paper addresses the implementation of a dual mode nonlinear model predictive control strategy integrating a nonlinear model predictive control (NMPC) with a local linear control method to control the product concentration in a continuous stirred tank reactor (CSTR), where a single exothermic reaction is carried out. For safe reactor operation, a reactor temperature path constraint is also considered in the NMPC formulation. Two well-developed linear controllers: state feedback linearizing control (SFC) and proportional-integral-derivative control (PID) technique are studied within the dual mode control framework. An Extended Kalman Filter (EKF) is incorporated into the dual mode NMPC to on-line estimate the unmeasured concentration of substances in the CSTR as well as uncertain key model parameters using an available temperature measurement. It has been found via simulation studies that the proposed dual mode NMPC, using either SFC or PID controller, connected with the EKF provides satisfactory control performance for set point tracking problem. The robustness of the controller in the presence of parametric uncertainties is also enhanced by the inclusion of the EKF. Moreover, the on-line computational time required by the dual mode NMPC is substantially decreased compared to that of an original NMPC.