Researchers from the Postgraduate Program in Metrology for Quality and Innovation and from the Department of Electrical Engineering of PUC-Rio have been developing GMI (Giant Magnetoimpedance) magnetometers for a wide range of applications. This project looks for the best way to apply the high sensitivity GMI magnetometers in Non-Destructive Testing (NDT) of steel structures, in order to identify small material discontinuities (defects) by the MFL (Magnetic Flux Leakage) technique. When submitted to low intensity magnetic fields, the GMI sensors (GMI ribbons) greatly vary the magnitude, and also the phase, of their impedances. Consequently, it becomes possible to apply them in the identification of small defects, such as cracks and small losses of thickness in steel sheets, since, in these discontinuities, there is a magnetic flux leakage (MFL) captured by the sensors. In this work, it was developed a study based on computer simulations, using the ANSYS Maxwell 3D software, aiming at the identification of the characteristics of the magnetic field curves that, caused by cracks, excite the GMI transducers, for different heights of these sensors in relation to the sheet under inspection. The analysis and post-processing, in the MATLAB software, of data extracted from the model developed in ANSYS allowed not only the identification of the best dimensional and electromagnetic configuration, but also the optimal operation point over the previously known characterization curves of the GMI samples.