Quasi-static and impact compression experiments were conducted on EA4T steel using a Gleeble-3800 thermal simulation machine. The study aimed to investigate the complex microstructural evolution and thermal deformation behavior of EA4T steel under various experimental conditions, including temperatures ranging from 970 °C to 1170 °C and strain rates ranging from 0.01s−1 to 1s−1. To precisely elucidate these phenomena, we meticulously constructed a unified visco-plastic constitutive model using the internal variable methodology. The model's parameterization was achieved through the effective application of genetic algorithm optimization techniques. Rigorous validation of the model was performed by meticulously comparing its outputs with experimental data, including key metrics such as average grain size, recrystallized fraction, and effective flow stress. In addition,a comparative analysis with the improved Arrhenius model highlights the superior performance of the unified visco-plastic constitutive equation in capturing the intricate microstructural evolution and thermal deformation behavior exhibited by EA4T steel.