In this paper a non-linear model based control (NMBC) scheme for the continuous stirred tank reactor (CSTR) process is proposed. The proposed control law is a function of the model state(s) and the measured output variable(s). Performance comparison of the proposed NMBC scheme with that of non-linear model predictive control scheme (NMPC) is reported. From the extensive simulation studies on the CSTR process, it can be inferred that servo and regulatory performances of the proposed NMBC scheme is found to closely match the NMPC scheme, with the proposed control scheme being computationally less intensive. In addition, a non-linear model based control scheme using an unscented Kalman filter (UKF-NMBC) and an inferential control scheme (INF-UKF-NMBC) for the CSTR process affected by random disturbances and random errors in measurements are formulated. Performance comparisons of the UKF-NMBC with UKF based NMPC and INF-AUKF-NMBC with INF-AUKF-NMPC, highlighting offset free servo performance, rejection of unmeasured disturbance, and robustness to model-plant mismatch are reported.