In this work, tensile tests were carried out on the 316L foil with a thickness of 0.1mm to calibrate the Gurson–Tvergaard–Needleman (GTN) model. Two calibration methods were compared for calibrating the same test data. The common approach is to use iterative finite element simulations to fit the experimental load-displacement curve. A new strain-based approach, which uses strain data obtained by a Digital Image Correlation (DIC) system, is presented here. It uses a stress-return-mapping algorithm to reconstruct the fracture related state variable, i.e., the accumulated void volume fraction. Fracture is predicted when the critical void fraction is reached at the measured fracture moment. The study shows that the new approach gives similar fracture model parameters compared to the common method, but with significantly higher computational efficiency.