The metal forming technology is (mainly due to the automotive industry) one of the most dynamically developing branch of the engineering industry. Continuous effort to achieve the top technological level and car´s safety factor at keeping the low price level means necessity to still implement into the own production process the newest mathematical models of these technological processes. Thus these days represents utilization of the numerical simulations an essential part for the car shape lay-out design, for determination the basic technological operations and also e.g. for stamping tools shape optimization. Alongside such implementation of the newest materials into production reveals necessity to develop new and more precise computational models of materials deformation behavior as well as models designed for spring-back prediction. Nowadays, in the branch of the metal forming technologies, there are several truly top software among which also belongs software PAM-STAMP 2G. In this article is evaluated influence of the computational model on the numerical simulation accuracy by PAM-STAMP 2G at the spring-back prediction. For the deformation analysis it was chosen stainless sheet material DIN 1.4301 and for the spring-back prediction were used two anisotropic computational models termed as Hill-48 and Vegter in combination with the kinematic hardening model termed as YOSHIDA UEMORI. Accuracy of the measured results from the individual computational models is evaluated by the compliance of the carried out experiment and results from the numerical simulations. For the own experiment was chosen test where material is drawn over the drawbead and drawing edge.