AbstractSignificant economic benefits can be derived from reducing the variability of product properties around their specification targets in batch reactors. End‐use properties in batch reactors using dynamic models that relate the operating conditions to the end‐use properties are related through a static nonlinear model to the molecular properties of the product and are controlled directly. Their values are required to lie in a desired target region in the property space. The control of end‐use product properties is formulated as a nonlinear model predictive control problem, and an efficient numerical technique using successive linearizations is utilized for the solution. A parameter adaptive extended Kalman filter is used for state estimation of the molecular properties that are not measured directly. The emulsion polymerization of styrene is chosen as a case study, and an approximate mathematical model is developed. The model is detailed enough to predict the end‐use properties, which depend on both the molecular weight and particle‐size distributions of the product. The end‐use product properties, such as tensile strength and melt index, are controlled in desired target regions by adding a monomer and a chain transfer agent, as well as the flow rate of the coolant, as the manipulated variables.