With more and more disease genes being mapped and/or cloned, there is a growing interest in dating the age of underlying mutations. The knowledge of the age of mutation is important to finely map disease genes by linkage disequilibrium mapping. It would also help us understand the origin, evolution, and dispersion of the mutant disease genes. Despite increasing interests in dating disease mutations, the development of appropriate statistical methods is largely fragmentary, and there is a lack of systematic treatment of the topic. We propose two classes of methods for estimating the age of mutant allele at the disease locus based on linked marker data. Our methods can not handle only single-locus marker data, but also multi-locus marker data as well. Moreover, our methods can be used even when the location of the disease locus is unknown, and/or when there are mutations at the marker or disease locus. We show that some previous results are special cases of our methods. We also derive a recursive equation previously obtained by Serre et al. [Hum Genet 1990;84:449-454] and provide an explicit solution to the equation. To illustrate our methods, we applied them to two groups of data sets, one is cystic fibrosis data collected from several European populations, and the other is data on several genetic diseases (diastrophic dysplasia, progressive myoclonus epilepsy, congenital chloride diarrhea, and Batten disease) all collected from the Finnish population. The former data set allows us to trace the origin and dispersion of the most common mutation for cystic fibrosis. The latter provides an opportunity to examine whether all mutations for these diseases have the same age.