An innovative approach was introduced for the development of a AA6063 recrystallization model. This method incorporated a regression-based technique for the determination of material constants and introduced novel equations for assessing the grain size evolution. Calibration and validation of this methodology involved a combination of experimentally acquired microstructural data from the extrusion of three different AA6063 profiles and results from the simulation using the Qform Extrusion UK finite element code. The outcomes proved the agreement between experimental findings and numerical prediction of the microstructural evolution. The trend of the grain size variation based on different process parameters was accurately simulated, both after dynamic and static recrystallization, with an error of less than 25% in almost the whole sampling computations.