The use of gradiometric configuration allows the optimization of the signal-to-noise ratio in magnetic measurements. GMI sensors are used in the measurement of ultra weak magnetic fields, which are typically bellow the the environment interference levels, being necessary the use of gradiometers to ensure satisfactory measurements. It is interesting that the response of these gradiometric configurations present high sensibility and linearity in the operating region. Is is also important that the sensing elements used present homogenic behavior, since the gradiometers performance is significantly affected by the homogeneity of the sensing elements. The sensibility, linearity and homogenization of the sensors is affected by the excitation current s characteristics (DC level, amplitude and frequency) and the polarizing magnetic field. Taking that into account, this work focuses in the implementation of a system based in genetic algorithms destined to the combination of the parameters responsible for optmizing the performance characteristics highlighted.