One of the primary goals of seismic design codes is to prevent structural collapse caused by large displacements resulting from severe earthquake ground motions. In general, using the force-based design method prescribed in seismic codes is a common practice for seismic design of structures. In design of structures using force-based methods applied in current seismic codes, to obtain the nonlinear displacements of structures under the design earthquake, the deflection amplification factor (Cd) is applied. In other words, the displacements obtained from elastic analyses under the reduced seismic forces are amplified by Cd to obtain the inelastic displacements under the design earthquake. Research studies showed that using a constant coefficient for estimating the inelastic displacements may lead to considerable overestimation or underestimation of the displacements in different stories of structures. The seismic behavior of a structure is under the influence of three important structural features, including stiffness, strength, and mass. Any change in these parameters would change the dynamic characteristics of the structure. The change is occasionally caused by a sudden change in the geometry of the structure along the height, such that the lateral dimension of the structure decreases significantly along its height at specific level(s). Such structures are named setback buildings. This architectural form enables the lower floors to gain sufficient sunlight when the adjacent structures are close together. A setback building has a sudden discontinuity in the frame geometry along the height. Generally, in regular structures, the inelastic maximum interstory drift ratio (IMIDR) occurs in lower stories. Investigating the seismic performance of structures with irregularity in their heights showed that the inelastic responses of these types of structures can differ significantly from the inelastic responses of regular structures. The present study investigates Cd for estimating IMIDR and inelastic maximum roof drift ratio (IMRDR) for steel special moment resisting frames (SMRFs) with setback, under the design earthquake. In addition, the variation of Cd with the variation of the location of the setback in the structural height and bays is studied. For this purpose, 24 steel SMRFs, including two groups of 5-and 9-story two-dimensional structures as low-and mid-rise structures, respectively, are investigated. Each group consists of one regular and 11 irregular setback structures. The structures are designed according to the Iranian Code of Practice for Seismic Resistant Design of Buildings (Standard No. 2800) and Specifications for Design and Construction of Steel Buildings (2022). The structures are designed using elastic modal response spectrum analysis. To perform nonlinear dynamic analyses, seven ground motion records that are matched to the design response spectrum are applied. To obtain Cd from the nonlinear analyses, the mean of IMIDR for each story is divided by the corresponding value obtained from the modal response spectrum analysis. Therefore, for each story of each structure, the Cd value is obtained. To calculate Cd for predicting real nonlinear roof displacement in the structures, the mean of the maximum of the roof displacement obtained from the nonlinear dynamic analyses is divided by the corresponding value obtained from the modal response spectrum analysis. The results show that for the 5-story setback structures (as low-rise setback structures), the recommended value of Cd = 5. 5 by Standard No. 2800 significantly underestimates the inelastic IMIDR values under design earthquake. Whereas, in the case of the 9-story setback structures (as mid-rise setback structures) the recommended value of the design code overestimates the inelastic IMIDR values in most of the cases. In addition, it is shown that the recommended value of Cd = 5. 5 by Standard No. 2800 can precisely predict the inelastic roof displacements of the setback structures under the design earthquake.