Abstract To know the Ancient and Medieval events and the recurrence interval of the great Kanto earthquakes such as the 1923 and 1703 ones generated by the subduction of the Philippine Sea plate along the Sagami trough, central Japan, I examined the latest dataset of historical records. I used only rank-A (contemporary) materials in the Online Database of Historical Documents on Japanese Earthquakes and Eruptions in the Ancient and Medieval Ages. Among destructive earthquakes in the Kanto district before the sixteenth century, the 878 Gangyo earthquake, which has been suspected an inland event generated from the Isehara fault, is considered the oldest candidate of the Kanto earthquake based on the large-scale disaster and intense aftershock activity, though tsunami is not mentioned. The 1293 Showo (Einin) Kamakura earthquake can be regarded as a great Kanto earthquake, because of severe damage in the Kanto district and remarkable aftershock activity. During the fifteenth century, both of the 1433 Eikyo and the 1495 Meio earthquakes are regarded as candidates of the Kanto earthquake; the former caused severe damage around Kamakura on the northern coast of Sagami Bay with high-aftershock activity and presumably caused tsunami, and the latter seems to have caused large tsunamis at Kamakura and the west coast of Sagami Bay. Although further investigation is necessary to clarify which one is the Kanto earthquake, we can say that an interplate earthquake probably occurred in the fifteenth century. Intervals between successive events, in 878, 1293, 1433 or 1495, 1703, and 1923, range from 140 to 270 yr except for the first interval of 415 yr. It is very difficult to confirm whether a great interplate earthquake occurred or not during this interval by means of historiographical seismology, because the Kanto district was in the worst situation of poor historical records in those days.