Due to the increasing demand for lightweight design in the automotive industry, twinning induced plasticity (TWIP) steels are now receiving extended attention, as they combine outstanding work-hardening potential with great plasticity. For the application of TWIP steels in car body parts, a precise description of the forming behavior under multiaxial loading is important for design purposes. Here forming limit curves (FLC) are an important tool to describe the failure behavior and are therefore an important input information for forming simulations. The ISO 12004-2 standard suggests using the position depending cross section method where a best-fit parabola is calculated based on staging points aside from the necking area. Due to the inhomogeneous strain distribution found in TWIP steels during multiaxial loading, which may lead to multiple necking areas and non-symmetric strain distribution, this method however has been found inapplicable. In the presented study, a novel approach to describe the forming limit of advanced high strength alloys suggested by Volk and Hora is applied on TWIP steel. In this line-fit approach, the temporal change of the thickness reduction rate is taken into account to identify areas of stable and instable deformation clearly and independent of the position and shape of the crack. It is found that the parabola-fit of the cross section method cannot reproduce the experimental results adequately when inhomogeneous strain distribution is observed, while the line-fit method is able to give stable and reproducible values for the whole FLC. On the other hand, it is found that a higher local resolution of the strain measurement leads to higher values for the FLC in case of the line-fit method while the cross section method is less sensitive to a change of strain measurement resolution.