Carbon steel is an important component in reinforced concrete structures. These composite materials are susceptible to various degradation processes, especially the corrosion of steel bars. Since they play an important role in the stability of reinforced concrete structures, these structures require periodic inspections to verify their physical conditions. Therefore, a nondestructive testing technique is proposed that is based on electromagnetic induction, which is capable of discriminating the diameters of steel bars. This technique was simulated with the COMSOL Multiphysics software and experimentally tested in the laboratory. Magnetic hysteresis loops were obtained for steel bars with 5.0, 6.3, 8.0, and 10.0 mm diameters at 0, 10, and 20 mm depths for sinusoidal and triangular excitation fields. The results obtained from the hysteresis loops for the maximum values of the field, remanence, and coercivity were processed by principal component analysis, Karhunen–Loéve transformation, Gaussian classifier, and artificial neural network. The confusion matrices produced by these pattern recognition techniques were analyzed, and the results show that the proposed method can distinguish different steel bar diameters, but is limited to a maximum depth.