There is still a great unawareness about the deep-water circulation in the NE-Central Atlantic since the Late Cretaceous. The morphology, sedimentary stacking pattern and distribution of contourite depositional systems have been widely used as paleoceanographic indicators, given clues about the relative velocity and pathways of bottom currents past-circulation. We present here evidence of a newcontourite features developed in the NE-Central Atlantic offshore Madeira Island between ~3000 m and ~ 4950 m water depth. These features allowed us to define two Contourite Depositional Systems (CDSs). The dataset used in our work iscomposed of multichannel reflection seismic profiles and DSDP and ODP sites for the chronostratigraphic framework. The seismic stratigraphy analysis allowed the identification ofsix seismic units (U1–U6) and a sub-unit (U6a) separated by erosional discontinuities (D1–D6). The acoustic basement corresponds to oceanic crust located in the Cretaceous Magnetic Quite Zone (~120 Ma to 84 Ma), but basalts drilled at DSDP Site 136 near Madeira Island indicate an age of about 106 Ma. Sedimentary deposits from seismic units are interpreted as pelagic (unit U1; Early Cretaceous); contourite (U2 to U6; Late Cretaceous to Quaternary); and mass transport deposits (U6a; Quaternary). These seismic units characterize the CDSs from Late Cretaceous (Campanian?), resting unconformably on pelagic sediments of probable Aptian to Santonian age. The CDS-1 (U2 to U4) was formed from Late Cretaceous (Campanian?) to Middle Miocene and the CDS-2 (U5 and U6) was deposited from Middle Miocene to Quaternary. They are bounded by major erosional surfaces and characterized by giant elongated mounded contourite drifts (Drift 1 and Drift 2, respectively). Presently, CDS-1 is inactive and buried by pelagic sediments. Conversely, CDS-2 outcrops between~3000 and 4800 m waterdepth on the lower slope of Madeira plateau. These results reveal that this region of the NE-Central Atlantic has been swept by long-term northward bottom currents. We propose that the most plausible water masses responsible for these bottom-currents and drifts generation have been the Southern Component Water (SCW) and more recently the Antarctic Bottom Water (AABW). Consequently, the onset of the first incursion of the SCW was near the end of Late Cretaceous, probably in the Campanian. Since the late Eocene the AABW would have a dominant role in this region of the NE-Central Atlantic. Assuming that during cool periods, the AABW had a greater volume and circulated shallower than 4000 m, we hypothesized that CDS-2 has been mainly active during these periods. On the contrary, during warmer periods, the AABW circulates deeper (>4000 m) and thus contourite deposition associated with the AABW shifted to deeper water depths. Presently the CDS-2 can be considered as a relict feature. Therefore, the Madeira CDCs represents a unique and very promising sedimentary archive for reconstructing deep-water masses circulation and their variability in the NE-Central Atlantic since the end of Cretaceous through the Quaternary, where the oceanic seafloor irregularities have been key in controlling the water masses and bottom currents behavior.