Mature ovarian teratoma is considered to be a parthenogenetic tumor that arises from a single oocyte/ovum. Conversely, complete hydatidiform mole (CHM) is androgenetic in origin: classic CHM arises from a single or two sperm. Since mature ovarian teratoma and CHM have only maternal and paternal genomes, respectively, their genome imprinting is theoretically reverse, but this has yet to be investigated. Genome imprinting in struma ovarii, a special form of mature teratoma, remains unclear. Although a mature teratoma can rarely arise in extragonadal sites, its genome imprinting, as well as cell origin, is poorly understood. One of the most important mechanisms of genome imprinting is DNA methylation. To investigate the methylation profile of imprinted genes, we performed methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA) of 21 imprinting control region (ICRs) of 9 imprinted genes/gene clusters in formalin-fixed, paraffin-embedded samples obtained from 12 mature ovarian teratomas, 6 struma ovarii, 10 CHMs, and 7 extragonadal (1 sacrococcygeal, 6 mediastinal) mature teratomas of females. In mature ovarian teratomas, ICRs of maternally and paternally imprinted genes showed high and low levels of methylation, respectively, and this pattern was almost reverse in CHMs. In CHMs, however, some ICRs showed aberrant methylation. The methylation profile of struma ovarii was comparable to that of mature ovarian teratomas, except for an adenomatous tumor. In extragonadal mature teratomas, the methylation pattern was somatic or irregular. In conclusion, mature ovarian teratomas/struma ovarii, CHMs, and extragonadal mature teratomas showed distinct methylation profiles of imprinted genes. Ovarian teratomas and CHMs are most likely to inherit their methylation profiles from their ancestral germ cells, although some aberrant methylation suggests a relaxation of imprinting in CHMs and a subset of struma ovarii. Extragonadal mature teratomas may carry a methylation profile of misplaced primordial germ cells or possibly somatic cells that have been reprogrammed in vivo.