Susceptibility to multiple sclerosis (MS) is determined by both inherited and non-inherited factors. The importance of genetic factors is demonstrated by the increased risk of disease in relatives of MS patients. Our objective was to determine the implications of the observed familial recurrence risks for the genetic basis of MS. We developed a computer program which calculates recurrence risks for monozygotic (MZ) twins, siblings, and second degree relatives, and used it to calculate recurrence risks for a wide variety of genetic models. We investigated models with different numbers of genes, different patterns of interaction between the genes, and dominant or recessive inheritance. The models that best reproduced the observed values had multiple loci with strongly synergistic interaction and autosomal dominant (AD) inheritance. At least six loci were required, and we found no upper limit on the number of loci. Models with genetic heterogeneity, where only a fraction of the risk loci are required for disease, are possible. In models with large numbers of loci the "abnormal" alleles conferring risk of disease are the most common allele. We conclude that a variety of genetic models with multiple genes, dominant inheritance, and synergistic interaction between risk genes are consistent with the observed familial recurrence rates in MS.