Batch and semi-batch reactors are used in the manufacture of low-volume high-value chemicals, where even a marginal increase in product yield can lead to considerably higher profits. Most of the present optimization and model based control schemes for semi-batch reactors are based on mechanistic models, whose development is a difficult and time- consuming exercise. Therefore, this book provides a new data driven approach for modeling, trajectory optimization and tracking of semi-batch reactors based on parameterization of input and output trajectories using orthonormal polynomials, and development of an artificial neural network model relating them using information available in historical databases. The generated optimal set point trajectories are tracked by developing data driven versions of generic model control. Simulation studies on semi-batch reactors illustrate their applicability. This can be useful to academic researchers working on data-driven modeling and optimization of batch processes and to industrial readers to explore the possibility of achieving better operational policies based on their historic operating data.