In this project, the Institute of Forming Technology and Lightweight Components together with Forvia-Faurecia are investigating the predictability of flow curves using artifical intelligence. Various approaches and strategies are used to solve the problems with regression, which are implemented using the Random Forest, Feed Forward Neural Network and Recurrent Neural Network models. To train the models, flow curves of different steels as well as parameters such as the yield strength and the tensile strength are used. In order to investigate the influence of the amount, scatter and noise of the available data on the prediction quality and robustness of the different approaches and models, a synthetic data set is created by a perturbation of existing flow curves. In the application, the trained model can be used to generate a yield curve from generic or batch-related characteristic values, which can be used, for example, within an FEM calculation.