Abstract: Objective To address complex residual stress distribution and associated risks of structural deformation and performance degradation in CMT additive manufacturing of aluminum alloy, this study investigates characteristics of residual stress and influence of welding parameters, providing a theoretical basis for process optimization. Methods A three-dimensional numerical model of residual stress was established based on thermophysical and mechanical properties of aluminum alloy and welding parameters. The model’s effectiveness was verified through comparison with residual stress measurements, and effects of different additive manufacturing speeds on residual stress distribution were systematically analyzed. Results Radial residual stress of the structure was relatively smaller and compressive. Circumferential residual stress was larger, with compressive stress dominant at inner wall and tensile stress at outer wall. Axial residual stress showed significant tensile stress on the outer side of middle section, while compressive stress appeared on the inner side. Equivalent residual stress was uniformly distributed along the thickness direction, with peak values near the bottom and middle regions approaching 120 MPa, close to material’s yield strength. Additive manufacturing speed had little effect on radial residual stress, but the mean circumferential tensile stress increased from 40 MPa to 45 MPa when the speed rose from 6 mm/s to 12 mm/s. Axial residual stress exceeded yield strength at 9 mm/s and then decreased from 130 MPa to 110 MPa. Overall, the peak residual stress increased with additive manufacturing speed, and high speed easily led to yield strength being reached. Compressive stress was consistently dominant at the inner wall, while tensile stress dominated at the outer wall. Conclusion The established model can effectively predict residual stress distribution during CMT additive manufacturing of aluminum alloy. By adjusting additive manufacturing speed, residual stress distribution can be optimized, reducing the risk of material yield failure and providing theoretical support for process optimization and forming quality improvement.