Beams are slender structures characterized by axial dimensions significantly exceeding their crosssectional dimensions. This geometric feature enables the representation of 3D beam models by equivalent 1D formulations, as demonstrated in Saint Venant's problem for isotropic materials with regular cross-sections in linear elasticity. For complex built-up beams, numerous dimensional reduction methodologies from 3D to 1D have been proposed. However, existing approaches remain confined to the geometrically linear regime, relying on kinematic relationships and equilibrium analyses in the undeformed configuration. In practical engineering scenarios, beams frequently undergo geometrically nonlinear deformations involving large displacements and rotations. Under such conditions, both the curvature of the beam's axis and the cross-sectional shape may deviate significantly from their undeformed states, necessitating advanced nonlinear modeling frameworks. This paper presents a novel center manifold-based approach for reducing three-dimensional beam models to one-dimensional models [1]. The reduced one-dimensional models consist of six equilibrium equations for the stress resultants and six constitutive equations that relate the derivatives of rigid-section motions to the stress resultants. These reduced models retain the accuracy of the full three-dimensional analysis while significantly improving computational efficiency. The key idea is to recognize that the Saint-Venant solution lies within a twelve-dimensional center manifold. In this approach, the beam's kinematics are decomposed into rigid-section motion and a warping field. The center manifold is parameterized by kinematic variables representing the six rigid-section motions and the six stress resultants on the cross-section. The nonlinear relationship between the warping field and the stress resultants, as well as the beam's nonlinear constitutive relationship, are approximated using polynomial expansions. The polynomial coefficients are determined by solving the invariance equation derived from the center manifold reduction process [2, 3]. By increasing the expansion order, the proposed method captures all nonlinearities present in the three-dimensional model and incorporates them into the one-dimensional beam's nonlinear equilibrium and constitutive equations. Numerical examples demonstrate both the efficiency and accuracy of the approach.