Abstract: The whole process true stress-strain relationship of steel material should be accurately considered when simulating the large deformation or fracture of the connecting parts or joint domains in steel structures. Firstly, uniaxial tensile tests on Q355 and Q420 steels were carried out to obtain the whole process engineering stress-strain curves. The finite element method was used to simulate the tensile tests of the steels. Through comparing the consistency between the finite element analysis and the measured engineering stress-strain curves, evaluated were the applicability and accuracy of Ling method, of Swift+Voce method, and of the improved Ramberg-Osgood method proposed for describing the post-necking true stress-strain curve of steels. To further explore the applicability of the three methods of describing the true stress-strain relationship for high-strength steels, the tensile tests on S690 and S960 steels reported in literature were simulated and analyzed. The results show that Ling method, Swift+Voce method, and the improved Ramberg-Osgood method proposed can accurately describe the whole process true stress-strain relationship of Q355 and Q420. Besides, these methods can also be used to describe the whole process true stress-strain relationship of S690 and S960. The universality of the three methods for describing the true stress-strain relationship of structural steels including high-strength steels is verified.