This work addresses the implementation of an online dynamic cascaded-conditional based optimization for handling uncertainty occurred in an autocatalytic esterification of Propionic Anhydride with 2-Butanol. The online optimization strategy includes an integration of the dynamic re-optimization mechanism (trigger, i.e. ±5% of conversion and dynamic re-optimizer, i.e. orthogonal collocation method in maximizing profit), estimator (cubature Kalman Filter) and controller (dual mode-adaptive PID). The re-optimization and control problems are solved separately in cascaded manner where the re-optimization mechanism is conditionally activated by using trigger. The simulation results show that the proposed strategy offers a large improvement in semi batch reactor performance if compared to the method which the optimal trajectories set point is pre-determined (offline optimization). Moreover, the online dynamic optimization of temperature and feed flowrate trajectories obtained was able to sustain the conversion within acceptable range (on-spec). Meanwhile, the offline optimization failed to handle the parameter model uncertainty that resulted to off-spec conversion and can lead to loss in profit.