Climate variability of the late Quaternary, especially the Last Glacial (LG) to the Holocene, has become the most heated topic for the recent decades, which helps to better understand the shape of current and future climate on our planet. The long term glacial/interglacial changes have been associated to insolation changes controlled by earth’s orbit, whereas the millennial scale variations are mostly accepted to be modulated by the “bipolar seesaw” mechanism which redistributes heat between the northern and southern hemispheres through the Atlantic Meridional Overturning Circulation (AMOC). The validation of such hypothesis is hampered by the very limited high resolution records from the high latitudes Southern Ocean. Situated at the southern end of the AMOC, Southern Ocean Atlantic sector represents one of the key regions for understanding the global climate change. The warm and cold water routes (WWR, south of Africa and CWR, Drake Passage/Scotia Sea) connect the South Atlantic to South Indian and Pacific Oceans, respectively; and the Weddell Gyre connects the open ocean South Atlantic to the Western Antarctic Shelf Ice (WASI), where nowadays the cold surface water and Antarctic Bottom Water (AABW) are generated beneath. These water masses represent the most important constituents of the AMOC in the Southern Hemisphere. This PhD project generated a series of new diatom based high resolution marine records covering wide area of the high latitudes South Atlantic, including from the Bouvet Island area and the Scotia Sea, aimed to provide new insights of the response and drive in Southern Ocean in the context of late Quaternary global climate change. With focusing on the LG to Holocene time period, by integration of our new generated and other existing records from the Southern Ocean Atlantic and Western Indian sectors, a detailed regional age model for the past 30 kyrs is established by AMS 14C dating and regional core correlation, which can be a template for further paleoenvironment reconstructions in this area. Our reconstructions suggest 2-3_C cooling in the LG south of the modern Polar Front compared to modern conditions. Winter sea ice in the Bouvet Island area expanded by 5_ latitude, the more expanded sea ice field resulted in the stronger tropical Atlantic cooling than other tropical oceans by the reduction of warm water to the South Atlantic via the WWR, and intensive export of carbon to the deep Southern Ocean. The two steps of deglacial warming are mainly concurrent with the Heinrich stadial 1 and Younger Dryas cooling in the northern hemisphere which support the bipolar seesaw from the marine archives. The North Atlantic Deep Water (NADW) provided the second source of warming from below to the high latitudes South Atlantic which extended the first step warming till 14 cal. ka (kyr BP). The temporal cooling at ca. 14-12 cal. ka is possibly caused by the melt water input from the Antarctica. Our records support the Southern Ocean control of the deglacial atmospheric CO2 rise. The early Holocene optimum marks the warmest period and the strongest cold water reduction and confined at ca. 11-10 cal. ka in the southern cores in the study area. Sea ice probably retreated south of modern conditions and maximum opal deposition southward expanded to at least 55_S. The mid-late Holocene cooling in the study area may be related to the cold water expansion from the Weddell Gyre with the developing cavity under the WASI. The Holocene climate development may have also interplayed with the NADW which represents a rapid resumption at the early Holocene and slight decline during the mid-late Holocene. The high resolution Holocene record from the central Scotia Sea suggests a stable climate at the core site. The early Holocene high productivity is maintained by the enhanced upwelling at that time. The Holocene reservoir change at the core site is evidenced, which is in agreement with the Southern II Abstract Ocean ventilation history. The pre-Holocene release of CO2 from the Southern Ocean strongly lowered the d14Catm which caused the low surface ocean reservoir at the early Holocene at the core site. The centennial scale climate variability may be mainly linked to solar activity and also influenced by the sea ice induced freshwater variability. Stratigraphis of long term records from the Scotia Sea covering the past 300 kyrs were established by a combination of radiocarbon chronology, correlation of magnetic susceptibility (MS) to Antarctic ice core dust/climate records, diatom biofluctuation stratigraphy, and geomagnetic chronology. Good consistency of the age models by these proxies improves the reliability of our stratigraphies and indicates the applicability of these approaches. However, detailed investigation show that, the radiocarbon chronology can be affected by the changes in carbon reservoir and fossil carbon contamination in the study area; the abundance pattern of diatom species Eucampia antarctica can be used to identify the past 6 marine isotope stages (MIS) while the fluctuation weakens during MIS 7 and 8; the reliability of geomagnetic chronology weakens at low sedimentation rates conditions. The MS-Antarctic ice core correlations represent high efficiency and the best reliability due to the high sensitivity to the changes of surrounding source regions and current systems, which is closely related to the climate changes in the Southern Ocean. In addition, a possible correlation between the ash layers found in our Scotia Sea cores and the Antarctic ice cores is established, which can be used as additional age markers for further studies in this area.