Atrial fibrillation (AF) is characterized by rapid and irregular contraction of the left atrium (LA). Impacting LA haemodynamics, this increases the risk of thrombi development and stroke. Flow conditions preceding stroke in these patients are not well defined, partly due the limited resolution of 4D flow magnetic resonance imaging (MRI). In this study, we combine a high-resolution computed tomography (CT) LA reconstruction with motion and pulmonary inflows from 4D flow MRI to create a novel multimodal computational fluid dynamics (CFD) model, applying it to five AF patients imaged in sinus rhythm (24 ± 39 days between acquisitions). The dynamic model was compared with a rigid wall equivalent and the main flow structures were validated with 4D flow MRI. Point-by-point absolute differences between the velocity fields showed moderate differences given the sensitivity to registration. The rigid wall model significantly underestimated LA time-averaged wall shear stress (TAWSS) (p = 0.02) and oscillatory shear index (OSI) (p = 0.02) compared to the morphing model. Similarly, in the left atrial appendage (LAA), TAWSS (p = 0.003) and OSI (p <
0.001) were further underestimated. The morphing model yielded a more accurate mitral valve waveform and showed low TAWSS and high OSI in the LAA, both associated with thrombus formation. We also observed a positive correlation between indexed LA volume and endothelial cell activation potential (ECAP) (R