Identification of forming limits of sheet metals is an important task to be done before the sheet metal forming processes. The information of the forming limit is indispensable for design of deformed shapes and related forming processes. This procedure becomes more important than ever as the auto-body becomes complicated and the number of auto-body parts is reduced for lower production cost. To identify the forming limit of sheet metals stretching with a hemispherical punch has gained popularity because of the convenient experimental procedure. The stretching experiment however has localized deformation or the shear band is originated from the non-unifrom deformation in the critical circum-stance instead of the absolute criterion. More accurate information of the forming limit therefore could be obtained by a more appropriate experiment to the real process. In this papaer an experiment program is devised to practivally identify the forming limits of sheet metals for auto-body parts. The experiment program contains not only stretching but deep-drawing Both forming experiments use the same hemispherical punch while they use different specimens. Deep-drawing experiments use speci-mens cut out in circular arc on both sides of circular blank to make it torn during the deep-drawing They also use speciments cut out straight in one side of a circular blank to make it deformed unevenly which causes local deformation during the deep-drawing. The experimental result demonstrates that the forming limit diagrams in the two cases show difference in their effective magnitude. The forming limit curve from deep-drawing is located lower than that from stretching. It is noted from the result that the deep-drawing process causes acceleration of localized deformation in comparison with the stretching process. From the experimental result the maximum value of forming limit could be pre-dicted for safe design.