Abstract: The accuracy of mechanical properties of sheet described by material model parameters has great influence on the accuracy prediction of springback. The material identifica- tion process can be time consuming if simulation model of experiment is complex. An inverse method, which is combined with FEM software ANSYS LS-DYNA and modified Levenb- erg Marguart(LM) optimization algorithm, is adopted to ident- ify the parameters in Barlat 1991 modal based on tensile test with rectangular specimens. It is time efficient and converges fast using this method to identify material parameters. There are two steps in the identification procedure. Firstly the Hill transversely anisotropic model is used and data from tensile test with specimen in rolling direction are adopted to calculate parameters in the hardening model. Secondly Barlat 1991 model is used and data from tensile test with specimens in 0°，45°，90° against rolling direction are adopted to calculate the parameters in the yield function. Results show that the parameters identified can predict the mechanical properties of the material relative accurately.

Key words: Anisotropic, Inverse identification Hardening model, Yield function