The electro-mechanical impedance (EMI) technique using piezoelectric lead zirconate titanate (PZT) transducers has been shown promising for health monitoring of steel structures. For damage assessment, statistical techniques such as root mean square deviation (RMSD) have been employed to compare the electro-mechanical (EM) admittance signatures acquired before and after damage. However, using the conventional RMSD approach, it is difficult to specify the damage severity and location. To overcome this limitation, a sub-frequency interval approach has been proposed by the authors for the concrete structures. In the present study, this approach is extended and applied to the steel structures. Considering low attenuation of wave and large sensing region of PZTs in steel structures, multiple PZTs are utilized to acquire the EMI signatures of an I-section steel beam. Artificial damage points are induced in the steel beam and the corresponding RMSD values for sub-frequency ranges (RMSD-S) are calculated. To qualitatively study the damage severity and location, a method combining the RMSD-S of multiple PZTs and the logarithmic attenuation of ultrasonic waves is proposed. The results demonstrate that the proposed method is robust and sensitive for damage assessment of steel structures.