In modern society, one of the most frequent clinical cases involves location and extraction of firearms projectiles, usually made of lead, a non-ferromagnetic material. The development of a technique that allows the precise location of these projectiles will aid their surgical removal, which has a great relevance because it contributes directly to the increase of the survival rate of wounded patients. This work presents and discusses a new approach, based on computational intelligence techniques, aiming at locating firearm projectiles inserted into the human body, by processing the information contained in magnetic field maps. The projectile is modelled by a sphere with radius a, located on a search space contained in a xy plane, that is situated at a distance h from the sensor, along the z axis. The x and y positions of the projectile were estimated by a windowing algorithm. In turn, the distance h between the sphere and the sensor is inferred by neural network, and the radius of the sphere a is estimated by a genetic algorithm. The obtained results indicate that the developed technique was capable to estimate all parameters with high accuracy and precision.