The main upwelling areas in Brazil are along its central coast. Among these, the Cabo Frio (23 deg S x 42 deg W) zone has the strongest signal of low sea surface temperature. The Cabo Frio coastal upwelling cell is controlled by three factors: the topography and orientation of the coastline; the position of the Brazilian Current axes, and the wind pattern, the last being the determining factor. Upwelling events in this region are concentrated in the austral spring and summer (September to April). Their occurrence is associated with NE-ENE winds, and they are inhibited when the wind blows from the southern quadrant (atmospheric frontal passages linked to polar advection). This pattern is associated with the seasonal displacements of the South Atlantic Convergence Zone (SACZ) over this region. Better knowledge of the SACZ behaviour over the time is very important to understanding the climatic variability over the South America in the Holocene. A high resolution sampling study is being done on two Kullenberg piston cores collected from the Brazilian continental shelf, 30-35 km southwest from the Cabo Frio island (115 m and 124 m water depth). AMS radiocarbon dating of organic matter from the core bases gives ages of 3,300 BP and 12,600 BP, respectively. We present the results of elemental analysis (TC and TN), Rock-Eval analysis (TOC, HI and OI), sedimentary organic matter petrographic analysis and alkenone sea surface paleo temperatures. Despite coming from an upwelling area, sediments have very low TOC content and show an increase from around 8,000 year with variations modulated by millennial and secular cycles. Sedimentary organic matter is marine type II that, associated with optical analysis, indicates a high-degraded state and a little continental contribution. The range of paleo temperature values are similar to results of other studies carried out in the same system and are coherent with regional Brazilian Current dynamics. These first results identify two distinct frequence modes of paleo temperature variability. The first one consists of long period variations, between about 12,000 and 6,000 years, that seems to be dependent on global climatic and oceanographic conditions. It reflects changes in the duration and intensity between NE vs. SW winds. The second is short period variations of about 6,000 years, which likely is linked to SACZ seasonal positioning that is controlled by interconnections between Pacific and Atlantic variability.