Marine biological carbon pump in the Southern Ocean is frequently referred as an important driver of atmospheric carbon dioxide concentration changes in the late quaternary world, but its variation pattern in the Indian sector remains elusive. We present micropaleontological and geochemical results during the last 800,000 years at site MD04-2718 (South Indian Ocean) to shed light on this issue. A total of 253 samples were analysed for coccolith assemblage at GEOPS, Paris Saclay University. For each sample, coccolith slides were prepared using the settling method of Beaufort et al. (2014; doi: 10.1038/nprot.2014.028), and subsequently analysed using the SYRACO system (Beaufort et al.,2014; doi: 10.1038/nprot.2014.028). The species composition, absolute abundance, and mean mass are reported. Calcite mass was also calculated following Brandon et al. (2022; doi: 10.1016/j.quascirev.2022.107556). A total of 96 samples were analysed for planktonic foraminifera assemblage at LSCE following the method of Brandon et al. (2022; doi: 10.1016/j.quascirev.2022.107556). Calcite mass of planktonic foraminifera in each sample was also estimated based on mean weights of the major planktonic foraminifera species from the fraction above 150 µm and their abundances. Besides, summer sea surface temperatures were estimated for each sample using the Modern Analog Technique of Therón et al. (2004; doi: 10.2113/50.4.391) and the transfer function of Govin et al. (2009; doi: 10.1029/2008PA001603).Organic matter contents were analysed on 230 samples at LSCE using a Flash EA 1112 elementary analyzer (Thermo Scientific). To this end, each sample was digested with ultra-pure HCl acid to remove carbonate. In addition, 66 samples were also selected for analysing their organic matter carbon isotope compositions using ThermoFisher Delta+XP IRMS at LSCE.Carbonate contents were analysed on 99 samples using the vacuum-gasometric technique described in Duchamp-Alphonse et al. (2018; doi: 10.1038/s41467-018-04625-7). In parallel, calcium X-ray fluorescence core scanner data (CaXRF) from Brandon et al. (2022; doi: 10.1016/j.quascirev.2022.107556) were quantitatively calibrated into Ca concentrations at the LSCE using an Energy-Dispersive X-Ray Fluorescence spectrometer (EDP-XRF, Epsilon3 XLE, Malvern PANalytical). The multivariate log-ratio calibration method of Weltje et al. (2015; doi: 10.1007/978-94-017-9849-5_21) was applied to calibrate the CaXRF signal intensities into high-resolution Ca concentrations (CaEDP-XRF) with a standard deviation of ± 0.3%. A high-resolution signal of CaCO3 content (CaCO3calculated) was obtained based on the linear correlation between measured CaCO3% and CaEDP-XRF. We also report new stable carbon and oxygen isotope results obtained from 365 samples of N. pachyderma s. and 263 samples of C. wuellerstorfi on this core, using a GV ISOPRIME isotope ratio mass spectrometer (IRMS) coupled with a carbonate extraction line at LSCE. Finally, the age model of the first 1915.44 cm of MD04-2718 are reported, which were established by graphical comparison between its planktonic foraminifera oxygen isotope record, sea surface temperature record, and calcium xrf signal with the deuterium isotope record from Antarctic ice cores.