Complex human diseases are often due to multiple disease genes with variable age of onset, as well as both genetic and environmental risk factors. In order to simultaneously consider two possible disease loci and to incorporate age of onset and covariate information into genetic linkage analysis, we introduce a two-locus genetic gamma frailty model for age of onset data within a family. We show that this model can be used for genetic linkage analysis of diseases due to two possible disease loci. This model simultaneously considers allele-sharing at two different loci, and can easily incorporate age of onset or age at censoring data, environmental covariates data, and population disease rate data in genetic linkage analysis. It also provides a flexible framework for mapping loci affecting the risk of developing diseases with complex etiology. This approach allows multipoint allele-sharing analysis to take evidence of the existence of a disease locus in one region into account in mapping other disease loci. Simulation studies indicate that the proposed tests have expected type 1 error rates and when the model assumptions are met, increased power over one-locus model for linkage analysis. A prostate cancer data set is used for illustrating the methods.