Cerebral aneurysms are an enlargement of a brain blood‐vessel due to a weakened wall and can pose significant health risks. Computational simulations have been thus utilized to help doctors in the understanding of cerebral aneurysms. In order to provide more accurate patient‐specific simulations, not only does geometry for the fluid domain need to be created from medical image data, but more accurate wall‐models need to be generated as well. However, the wall‐thickness and material properties are very difficult to obtain experimentally, and thus most computational simulations, except for a few we have seen in the recent years, are performed using walls with uniform thickness and constant material properties. To provide a more accurate computational model for the weakened wall‐structure, this paper presents a novel estimation of an equivalent cerebral aneurysm wall‐thickness by deforming a healthy vessel onto an aneurysm through surface parameterization and a non‐linear spring system. The resulting wall‐model is thinnest over the dome of the aneurysm, and for the patient‐specific models used has an average thickness of 75µm. Fluid–structure interaction simulations of three patient‐specific cerebral aneurysms are carried out. Compared with using a uniform wall‐thickness, using the equivalent wall‐thickness gives a more accurate prediction of the rupture site. Copyright © 2010 John Wiley & Sons, Ltd.