The Sancy (16 km(2)) is the youngest of the two stratovolcanoes that constitute the Mont-Dore Massif (Massif Central, France). The restricted number of high precision radio-isotopic ages currently limits our knowledge of the pyroclastic chronology of this edifice which is the source of many tephra layers detected in middle Pleistocene sequences in southeast Europe. To improve our knowledge of the building phases of this stratovolcano, we collected thirteen pyroclastic units covering the entire proximal record. We present Ar-40/Ar-39 single grain laser dating performed in the facility hosted at the LSCE (Gif-sur-Yvette, France). The Ar-40/Ar-39 ages range from 1101 +/- 11 ka to 392 +/- 7 ka (his external). Four pyroclastic cycles lasting on average 100 ka were identified (C. I to C. IV). C. I corresponds to the earlier explosive phase between 1101 ka and 1000 ka and starts about 100 ka earlier than previously thought. The second pyroclastic cycle (C. II) is the main pyroclastic episode spanning from 818 to 685 ka. This cycle is constituted of a minimum of 8 major pyroclastic eruptions and includes a major event that corresponds to a large plinian eruption at 719 +/- 10 ka (1 sigma external) and recorded as a 1.4 m thick layer 60 km south-east of the Sancy volcano. The link between this large eruption and formation of a caldera stays however, hypothetical. The third pyroclastic cycle (C. III) found in the northeastern part of the Sancy (Mont-Dore valley) spanned from 642 to 537 ka. Finally, the youngest pyroclastic cycle (C. IV) starts at 392 ka and probably ends around 280 ka. The age versus geographic location of each pyroclastic cycle indicates three preferential directions of channeling of the pyroclastic events and/or collapse of the volcanic edifice: northwest to west (C. l), southeast (C. II) and finally north to northeast (C. III and IV). The new high precision Ar-40/Ar-39 age for the Queureuilh bas pyroclastic unit (642 +/- 9 ka) is identical within error with the U/Pb age obtained by Cocherie et al. (2009) [Geochimica et Cosmochimica Acta, 73, 1095-1108] and suggests a short residence time of the magma in a shallow, short-lived, small magmatic chamber. Finally, the source of the t21d tephra layer found in the Pianico Sellere varved sequence (Northern Italy) is not the Rivaux pumice flow as proposed by Brauer et al. (2007) [Journal of Quaternary Science 22, 85-96] and neither one of the C. II pyroclastic units as suggested by Roulleau et al. (2009) [Quaternary International 204,31-43]. Accordingly, the source for the t21d layer has yet to be found at Sancy or elsewhere.