Nonchromosomal pregnancy failure is a common but poorly understood phenomenon. Because recent data have suggested that epigenetic abnormalities such as abnormal placental DNA methylation may play a role in human pregnancy failure, we undertook experiments to test whether decidual and/or placental DNA methylation abnormalities are present in a mouse model of pregnancy failure. A large number of studies have shown that crosses between CBA/J female mice and DBA/2 males result in pregnancies with a high rate of failure/resorption, whereas other crosses with CBA/J females produce normal pregnancies. Although the CBA/J × DBA/2 mouse has frequently been used as a model for miscarriage, a detailed explanation for the pregnancy failure phenotype is lacking. We performed timed matings between CBA/J female and DBA/2 male mice as well as between DBA/2 female and CBA/J male mice. Decidual caps were isolated at Embryonic Day (E) 9.5 from both crosses, and a microarray-based method was used to comparatively assess genomic methylation at approximately 16 000 loci on mouse chromosome 7. In comparison with decidual caps from DBA/2 × CBA/J pregnancies, CBA/J × DBA/2 decidual caps were characterized by widely and apparently randomly disturbed methylation. In another set of analogous experiments, genomic methylation of placental DNA from E8.5 pregnancies was assessed using the same microarray-based method. This analysis revealed that in contrast to the decidua, placental DNA methylation from CBA/J × DBA/2 pregnancies was indistinguishable from that of normal controls. We conclude that abnormal DNA methylation in the uterine decidua likely plays a role in the CBA/J × DBA/2 model of pregnancy failure. To our knowledge, these experiments are the first to demonstrate that epigenetic abnormalities of the decidua are associated with pregnancy failure, and they set the stage for future efforts to understand the role of DNA methylation at the maternal-fetal interface.