<p>The Mg-6.0 Zn-2.0 Yb-0.5 Zr (ZYbK620) alloy has attracted considerable attention owing to its tunable microstructure and excellent mechanical properties. Nevertheless, a predictive model accurately capturing dynamic recrystallization (DRX) and microstructural development during hot deformation has yet to be established. In the present study, isothermal hot-compression tests were conducted on a solution-treated ZYbK620 alloy using a thermomechanical simulator at 250, 300, and 350°C and strain rates of 1, 0.1, 0.01, and 0.001&#xa0;s⁻¹. The constitutive model was calibrated using artificial neural networks (ANNs) based on stress-strain data to achieve high predictive precision. Subsequently, DRX kinetics and cellular automaton (CA) models were developed, and a coupled cellular automaton-finite element (CA-FE) model was constructed. Simulations were performed under the deformation condition of 325°C and 0.05&#xa0;s⁻¹, and the results were compared with experimental observations. The simulated DRX behavior showed good agreement with experimental measurements, confirming the predictive accuracy and robustness of the developed models. These findings provide a theoretical basis for microstructural optimization and the design of hot-working processes for the high-performance ZYbK620 alloy.</p>