Abstract: With the rapid development of global economy, problems in energy production and environmental protection are becoming severe, and the automotive industry is under increasing pressure to reduce the weight of vehicles and improve crash performance. Due to the demand for reduced vehicle weight as well as improved safety and crashworthiness, hot-stamped components from ultra-high strength steels have been utilized for automobile manufacturing. Currently, the most widely used hot-stamped steel plate is 22MnB5. Its tensile strength is 1500 MPa and yield strength is 1200 MPa. In contrast, as the demands for steel strength have increased, the demand for high strength grades of steel has been quickly put on the production agenda. In recent years, a novel hot-stamped steel, 38MnB5 has been developed, with a tensile strength exceeding 2000 MPa. The high temperature deformation behavior of 38MnB5 steel was investigated by the Gleeble-3500 thermal-mechanical simulator. The isothermal uniaxial tensile tests of the steel were performed within deformation temperature range of 650-950℃ under strain rates of 0. 01, 0. 1, 1, and 10 s^-1, and the typical true stress-strain curves of 38MnB5 at relative conditions were analyzed. The experimental results show that the flow stress rises with decreasing deformation temperature under the same strain rate, and with an increasing strain rate. When the strain rate gradually increased, dynamic recovery and dynamical recrystallization exhibited an apparent effect on the hot deformation process, while the inconspicuous impact receded with rising temperature. In consideration of the multiple influences on deformation temperature, strain rate and strain, a phenomenological, constitutive relationship was developed to depict the hot deformation process of 38MnB5. In the established equation, the material constants dependent on the deformation temperature, strain rate, and strain were obtained using regression analysis of the experimental data for flow stress, strain, strain rate, etc. The comparison between the calculated data and the experimental data show that the calculated data derived from the constitutive models are found to be in satisfactory agreement with the experimental results.