Fault-bend folding theory describes the geometric and kinematic evolution of structures in fold-thrust belts and is widely applied to characterize hydrocarbon systems and evaluate subsurface energy storage and carbon sequestration sites. The theory is solely kinematic and does not account for aspects of structural growth that are influenced by mechanical stratigraphy or stress. To address this, we established a mechanical basis for fault-bend folding using models developed with the discrete element method (DEM). The DEM models use an aggregate of circular, frictional disks that incorporate bonding at particle contacts to represent the numerical stratigraphy. Displacement of the hanging wall along a predefined fault leads to the development of an emergent fold as strata pass across the fault bend. We used this numerical sandbox setup to study the mechanics of fault-bend folding with varying mechanical strength, stratigraphic layering, and fault geometries. Both anticlinal and synclinal fault-bend fold cases produce well-defined fold limbs that generally reproduce the primary characteristics of kinematic fault-bend fold models. In the presence of mechanical layering, folds develop primarily by flexural slip; in the absence of layering, folds develop by shear along discrete faults that are generally parallel to axial surfaces. The pregrowth strata of synclinal fault-bend fold models accord very closely with the kinematic theory over a wide range of fault and fold geometries; however, the anticlinal DEM fault-bend folds exhibit behavior that is distinct from kinematic predictions. Specifically, we find that the anticlinal folds maintain a linear relationship between fold shape (Γ and γ) and cutoff angle (θ) for a given fault bend (ϕ). This behavior occurs over a wide range of model properties, both with and without mechanical layering, and is clearly distinct from the two modes of anticlinal fault-bend folding prescribed by the kinematic theory. The patterns of folding in growth strata are also distinguished from classic fault-bend fold models in that they exhibit an upward decrease of the bed dips within growth triangles. These differences arise because the DEM models incorporate a broader range of deformation mechanisms that operate across axial zones of finite width. These differences observed in the DEM models lead us to define a revised quantitative relationship between fold and fault shape that can serve as a useful interpretation tool.