The hemophilia A mutation database lists more than 160 missense mutations: each represents a molecular defect in the FVIII molecule, resulting in the X-linked bleeding disorder hemophilia A with a clinical presentation varying from mild to severe. Without a three-dimensional FVIII structure it is in most cases impossible to explain biological dysfunction in terms of the underlying molecular pathology. However, recently the crystal structure of the homologous human plasma copper-binding protein ceruloplasmin (hCp) has been solved, and the A domains of FVIII share approximately 34% sequence identity with hCp. This advance has enabled the building of a molecular model of the A domains of FVIII based on the sequence identity between the two proteins. The model allows exploration of predictions regarding the general features of the FVIII molecule, such as the binding-sites for factor IXa and activated protein C; it has also allowed the mapping of more than 30 selected mutations with known phenotype from the database, and the prediction of hypothetical links to dysfunction in all but a few cases. A computer-generated molecular model such as that reported here cannot substitute for a crystal structure. However, until such a structure for FVIII becomes available, the model represents a significant advance in modeling FVIII; it should prove a useful tool for exploiting the increasing amount of information in the hemophilia A mutation database, and for selecting appropriate targets for investigation of the structure-function relationships via mutagenesis and expression in vitro.