This study develops a forward-inverse scheme that combines the streamline method with the adaptive neuro-fuzzy inference systems (ANFIS) to determine the anisotropic optimum blank in cylindrical cup drawing process. After cylindrical cup drawing process, it is not easy to obtain the same cup height because of material anisotropy. The prediction scheme could efficiently improve the anisotropy of the sheet metals causes the ear phenomenon. First, the forward scheme of the streamline method could calculate the profile of the nearby optimum blank. Then, the basic database for the inverse scheme of ANFIS is established by explicit dynamic finite element method (FEM). Using this database, the exact optimum blank profile can be inversely predicted more correctly by hybrid-learning cycles of ANFIS. The deformed contours of the experiment and FEM simulated results are compared using the exact optimum blank. It shows that a good agreement is achieved. From this investigation, the forward-inverse scheme is proved to be able to supply a useful optimal soft computing process in the anisotropic sheet metal forming category.