The latest version of the Standard for Structural Calculation of Reinforced Concrete Structures, published by the Architectural Institute of Japan in 2010 [1], allows the design of shear walls with rectangular cross sections in addition to shear walls with boundary columns at the end regions (referred to here as “barbell shape”). In recent earthquakes, several reinforced concrete (RC) shear walls were damaged by flexural failures through concrete compression crushing accompanied with buckling of longitudinal reinforcement in the boundary areas. Damage levels have clearly been shown to be related to drift in structures; this is why drift limits are in place for structural design criteria. A crucial step in designing a structure to accommodate these drift limits is to model the ultimate drift capacity. Thus, in order to reduce damage from this failure mode, the ultimate drift capacity of RC shear walls needs to be estimated accurately. In this paper, a parametric study of the seismic behaviour of RC shear walls was conducted using a fibre-based model to investigate the influence of basic design parameters including concrete strength, volumetric ratio of transverse reinforcement in the confined area, axial load ratio and boundary column dimensions. This study focused on ultimate drift capacity for both shear walls with rectangular sections and shear walls with boundary columns. The fibre-based model was calibrated with experimental results of twenty eight tests on shear walls with confinement in the boundary regions. It was found that ultimate drift capacity is most sensitive to axial load ratio; increase of axial load deteriorated ultimate drift capacity dramatically. Two other secondary factors were: increased concrete strength slightly reduced ultimate drift capacity while increased shear reinforcement ratio and boundary column width improved ultimate drift capacity.