A window wall is a lightweight external wall composed of glass and aluminium, which does not carry any gravity loads from the building except for its own weight. Wind load is transferred to the main structural system through the connections of sub-frames (consisting of sub-heads and sub-sills) with slabs. An aluminium sub-head with a removable bead (ASHWRB) is a common type of sub-head consisting of two parts: a base and a bead. Under wind load, the bead flange is directly loaded by the mullion and undergoes bearing failure due to its long length. This paper presents a detailed numerical study on the bearing capacities of the ASHWRB sections subject to wind loading. For this purpose, finite element models were simulated using ABAQUS/CAE and validated with our previously conducted experimental work in terms of ultimate loads, load-displacement curves, and failure modes. Validated finite element models of the ASHWRB sections were then used for an extensive parametric study covering a wide range of thicknesses, flange widths, bearing widths, and engagement lengths. Notably, the ultimate loads obtained from the parametric study are not product nor manufacturer specific, whereas, the data obtained from experimental study is specific to the manufacturer products. The acquired bearing database was used for a detailed assessment of the consistency and reliability of the previously proposed design rules by the authors (DRA) based on the experimental tests as well as currently available cold-formed steel design rules (DRSs). It was found that the bearing capacities determined by both DRA and DRSS are unreliable for predicting the bearing capacities of the ASHWRB sections in the window wall. Thus, modifications were made to DRA and DRSS. Consequently, new design rules were proposed in this paper (DRAP), which accurately predicted the bearing capacities and correlated very well with the numerical results.