To explain the origin of a high heat flow anomaly observed within 150 km seaward of the Japan Trench, we construct a thermal model for an oceanic plate prior to subduction that includes the effect of hydrothermal circulation within a high-permeability aquifer in its uppermost part. The model includes the effects of aquifer thickening, which is expected to occur near subduction zones where plate bending prior to subduction causes fracturing and faulting within the oceanic plate. Using typical parameter values for the Japan Trench, we find that hydrothermal circulation in the thickening aquifer mines heat from the underlying basement and can account for the observed high heat flow anomaly. The ratio of heat supply below the aquifer as a result of aquifer thickening to the inverse of the thermal resistance of the sediment layer is a control parameter for the system. As long as the aquifer permeability is higher than ∼10−13 m2, a typical value for the uppermost part of the oceanic plate, variations in other details of the hydrothermal circulation such as the exact value of the aquifer permeability and the size of the convection cells do not significantly change model results. Despite its strong influence on seafloor heat flow seaward of the trench, this hydrothermal heat mining does not affect significantly the thermal structure of the subducted oceanic plate. This finding indicates that surface heat flow anomaly around the trench may not correspond to temperature anomaly within the subducted oceanic plate and the megathrust seismogenic zone.