Background and aims: The root reinforcement (RR) models commonly used in slope stability modelling can be simply explained as a single soil additional cohesion parameter estimated with simple analytical functions of root traits. We have simulated 3D direct shear tests using the standard implicit Finite Element Method (FEM) and the Discrete Element Method (DEM), aiming to (i) evaluate the RR models and (ii) compare the two numerical approaches. Methods: In homogeneous soil with low cohesion, 36 straight, non-branched and thin root models were implanted in three parallel lines. Root traits, including orientation relative to the shear direction (45°, 90° and −45°), longitudinal modulus of elasticity (10 MPa and 100 MPa), and bending and compressive behaviours (beam, truss and cable) were investigated. Results: Compared to the FEM, the DEM achieved consistent results and avoided convergence problems, but required longer computation time and used parameters potentially difficult to identify. Root reinforcement did not occur until significant plastic deformation of soil. The RR values estimated by the shear tests were much lower than those estimated by the usual RR models and were significantly dependent upon root traits. Conclusions: Ignoring the effect of root traits in RR models might lead to an important bias when using slope stability models.