The hardening behavior of metals, as affected by strain path changes, is one of the reasons preventing accurate sheet metal forming predictions. In particular, martensitic steel exhibits severe springback due to its very high strength and because of the hardening effects induced by strain path changes. This limitation can be mitigated by more extensive utilization of advanced constitutive models in the numerical optimization of a process. In the present work, the classical isotropic hardening model, the Yoshida-Uemori two surface kinematic hardening approach, and the homogeneous anisotropic distortional hardening framework are selected to conduct numerical simulations. Two forming applications, the C-rail and Roof-side-rail, are conducted to validate the three hardening approaches discussed above. The springback profile for each of these models is compared with the experimental results. For the two forming examples considered, the distortional hardening model appears to be the most suitable to predict the springback.