The role of hemodynamics and mechanobiology in thoracic aorta (TA) diseases is well-established. Computational Fluid Dynamics (CFDs) is a powerful method to study blood flow in cardiovascular districts. Commonly, CFD simulations rely on the rigid-wall assumption. Instead, Fluid-Structure Interaction simulations, while being more accurate, have high computational times and need additional information on vessel wall. Recent studies applied mesh morphing to cope with the ascending TA deformations, showing however some limitations. This work aims to develop a new image-based method for the entire TA (i) to set-up moving boundaries CFD simulations based on mesh morphing techniques and (ii) to get a mechanical characterization of the wall. Starting from 20-phase ECG-gated CT images of 10 patients, we built 3D models of the TA and of the left ventricle (LV) through a custom developed 3D U-net. Firstly, the TA 3D models were morphed on the baseline mesh (0%) by applying an in-house registration algorithm. The wall displacement was then employed to set-up moving-boundary CFD simulations (CFDm). Hemodynamic results were compared with those obtained from rigid-wall CFD simulations run on the baseline mesh (CFD0). Inlet flow conditions extracted from the LV volumes were set. Secondly, the mapped TA 3D models were employed to assess wall circumferential stress, strain and stiffness, also considering the aortic twist at the valve plane. The developed simulation strategy copes with TA morphological changes during the cardiac cycle. The CFDm method models the flow waveform shift occurring along the TA lumen and highlights differences in hemodynamic parameters, with respect to CFD0 approach, overcoming the limitations of state-of-the-art simulations. Moreover, the image-based tool allows to estimate stiffness and strain distributions for all the studied subjects, resulting in a promising method for non-invasive estimation of the aortic wall mechanical properties.