The rate of post-operative complications is the main drawback of endovascular repair, a technique used to treat abdominal aortic aneurysms. Complex anatomies, featuring short aortic necks and high vessel tortuosity for instance, have been proved likely prone to these complications. In this context, practitioners could benefit, at the preoperative planning stage, from a tool able to predict the post-operative position of the stent-graft, to validate their stent-graft sizing and anticipate potential complications. In consequence, the aim of this work is to prove the ability of a numerical simulation methodology to reproduce accurately the shapes of stent-grafts, with a challenging design, deployed inside tortuous aortic aneurysms. Stent-graft module samples were scanned by X-ray microtomography and subjected to mechanical tests to generate finite-element models. Two EVAR clinical cases were numerically reproduced by simulating stent-graft models deployment inside the tortuous arterial model generated from patient pre-operative scan. In the same manner, an in vitro stent-graft deployment in a rigid polymer phantom, generated by extracting the arterial geometry from the preoperative scan of a patient, was simulated to assess the influence of biomechanical environment unknowns in the in vivo case. Results were validated by comparing stent positions on simulations and post-operative scans. In all cases, simulation predicted stents deployed locations and shapes with an accuracy of a few millimetres. The good results obtained in the in vitro case validated the ability of the methodology to simulate stent-graft deployment in very tortuous arteries and led to think proper modelling of biomechanical environment could reduce the few local discrepancies found in the in vivo case. In conclusion, this study proved that our methodology can achieve accurate simulation of stent-graft deployed shape even in tortuous patient specific aortic aneurysms and may be potentially helpful to help practitioners plan their intervention.