Subduction zones are the loci of active seismicity and large tectonic deformation.The foot of the accretionary prism has long been considered as aseismic zones whereearthquake energy dissipates, but recently, mega-earthquakes have activated largedisplacements near subduction trenches, generating devastating tsunamis. It is important to better understand the seismotectonics of these zones. The Nankai subduction zone is one of the most suitable study sites: in addition to the seismic hazardof mega-earthquakes (M w > 8) for densely populated Japanese coasts, it is also thesite of slow slip. This was the motivation for the NanTroSEIZE project, a flagshipproject of the International Ocean Discovery Program (IODP), which consisted of13 expeditions from 2004 to 2019. A detailed understanding of the hydraulic regimeis needed to understand the diversity of behavior at Nankai. Indirect evidence sug-gests that this subduction zone is particularly fluid-rich. To refine these conclusions,we innovatively integrate drilling and logging data acquired in the NanTroSEIZEproject, in the frontal region (sites C0024 and C0006) and in the ”megasplay” faultregion (sites C0004 and C0001). We obtain a high-resolution hydraulic propertiesprofile along the NanTroSEIZE drillings, to finely characterize the fault architecturein the Nankai subduction zone. This will help to estimate: (a) pore pressure profilesacross the accretionary prism (b) high-resolution hydraulic profiles to identify thedamaged zone around the detachment fault, and verify whether it is a significantdrainage path. (c) permeability evolution. This allows inferring the origin of thefluids (diffusion along the décollement or local fluid emission following compactionof the sedimentary pile), and extrapolating the data near the pit to deeper seismically active zones. Mud pressure reanalysis allows us to estimate the inflow of fluidsfrom the formation to the borehole. To verify its predictions, 2 methods have beenused to predict pore pressure: (a) Eaton-type methods based on drilling or seismicdata (b) empirical relationships between pore pressure, porosity, and P-wave velocity, obtained from consolidation experiments. Permeabilities can be calculated bycombining flow and pressure estimates. At site C0024, a fluid influx (> 0.05m3/s)is localized in the damaged zones of the décollement at 813 mbsf. It is associatedwith a significant overpressure (P ∗≈2.38 − 4.79 MPa) below the plate boundary.The décollement acts as a vertical hydraulic barrier, but as a longitudinal drainthat diffuses overpressures from the site where earthquakes and slow slides occur.In contrast, at site C0006, fluid influxes remain weak, and localized just below the frontal thrust at 711 mbsf. There is no apparent increase in pore pressure within theaccretionary prism, perhaps because a fault between C0024 and C0006 drains theoverpressurized fluids. In the region of the megasplay fault at site C0001D, majorfluid flow (> 0.107m3/s) is observed along a fault zone below 500 mbsf. The observedmud pressure is supra-lithostatic. At site C0004A, a large fluid flow (0.07m3/s) andhigh pore pressure are also observed in the fault zone at about 310 m depth. Thelarge fluid flow is evidence of high pore pressure, facilitating propagation of the co-seismic slip to the seafloor. The megasplay region is then more favorable than thefoot of the prism region for generating tsunamigenic earthquakes.