Heat flow density (HFD) has been derived in the lower Congo basin from 293 conventional probe measurements, 424 oil exploration data, and about 1,000,000 Bottom Simulating Reflector (BSR) values. Each set of data provides independent estimates of temperature gradients associated with specific uncertainties and geographic distributions, but a statistical comparison in overlapping domains shows a relatively good consistency. In situ conductivity has been measured in conventional surveys and estimated from rock composition and porosity in oil exploration boreholes: in the deep offshore, measurements and estimates are consistently low and increase significantly toward the shelf domain. Thermal conductivity does not change significantly in the interval between the water bottom and the BSR and can be estimated from the surface measurements to provide an inference of HFD. The combined sets of HFD in the lower Congo basin cover the overall domain of the continental margin and a part of the oceanic domain. High-resolution imaging of the BSR by three-dimensional (3-D) seismic studies provides, in addition, good quality (nominal resolution of ~500 m) in the deep offshore. At short length scales (on the order of kilometers), large variations of HFD (from 30 to 140 mWm-2) exist above salt diapirs and canyons, related to heat refraction. On the other hand, no small-scale effect related to fluid flow has been detected, although active venting is known. At larger length scale, HFD in the deep offshore is different on both sides of the Congo submarine channel (54 ¿ 8 mWm-2 to the north and 48 ¿ 12 mWm-2 to the south), which are also characterized by different basin-tectonic styles. At the margin scale, HFD is significantly different in the oceanic domain (42 ¿ 3 mWm-2), the deep continental offshore (52 ¿ 10 mWm-2), and the continental shelf (65 ¿ 15 mWm-2). A progressive reduction of the radiogenic heat contribution in the extended continental crust is the more likely the cause of these differences. Mantle heat flow below the continental margin is consequently high compared to that expected (~15 mWm-2) for old shields such as Congo.