The variability of water availability in the Amazon Basin is highly related to the sea surface temperature (SST) pattern over the Atlantic and the Pacific Oceans. The possibility of forecasting this variability one season in advance is a necessity for the correct management of water in the region for water supply, energy production, and flood control. Here we present the development of a statistical model to forecast seasonal discharge, one season in advance, at 12 sites in northeastern South America from Pacific and Atlantic Ocean SSTs. Results show that the model has a better capacity to forecast discharge at some sites than at others and that different oceanic regions have different influences over the basins. Discharge at subbasins in the northern part of the Amazon Basin are better forecasted from Pacific Ocean SSTs while discharges at subbasins in the southern part are better forecasted from Atlantic Ocean SSTs. ¿ 1998 American Geophysical Union