The Cretaceous–Paleogene (K–Pg) mass extinction event 66 million years ago led to large changes to the global carbon cycle, primarily via a decrease in primary or export productivity of the oceans. However, the effects of this event and longer-term environmental changes during the Late Cretaceous on the global sulfur cycle are not well understood. We report new carbonate associated sulfate (CAS) sulfur isotope data derived from marine macrofossil shell material from a highly expanded high latitude Maastrichtian to Danian (69–65.5 Ma) succession located on Seymour Island, Antarctica. These data represent the highest resolution seawater sulfate record ever generated for this time interval, and are broadly in agreement with previous low-resolution estimates for the latest Cretaceous and Paleocene. A vigorous assessment of CAS preservation using sulfate oxygen, carbonate carbon and oxygen isotopes and trace element data, suggests factors affecting preservation of primary seawater CAS isotopes in ancient biogenic samples are complex, and not necessarily linked to the preservation of original carbonate mineralogy or chemistry. Primary data indicate a generally stable sulfur cycle in the early-mid Maastrichtian (69 Ma), with some fluctuations that could be related to increased pyrite burial during the ‘mid-Maastrichtian Event’. This is followed by an enigmatic +4‰ increase in δ34SCAS during the late Maastrichtian (68–66 Ma), culminating in a peak in values in the immediate aftermath of the K–Pg extinction which may be related to temporary development of oceanic anoxia in the aftermath of the Chicxulub bolide impact. There is no evidence of the direct influence of Deccan volcanism on the seawater sulfate isotopic record during the late Maastrichtian, nor of a direct influence by the Chicxulub impact itself. During the early Paleocene (magnetochron C29R) a prominent negative excursion in seawater δ34S of 3–4‰ suggests that a global decline in organic carbon burial related to collapse in export productivity, also impacted the sulfur cycle via a significant drop in pyrite burial. Box modelling suggests that to achieve an excursion of this magnitude, pyrite burial must be reduced by >15%, with a possible role for a short term increase in global weathering rates. Recovery of the sulfur cycle to pre-extinction values occurs at the same time (∼320 kyrs) as initial carbon cycle recovery globally. These recoveries are also contemporaneous with an initial increase in local alpha diversity of marine macrofossil faunas, suggesting biosphere-geosphere links during recovery from the mass extinction. Modelling further indicates that concentrations of sulfate in the oceans must have been 2 mM, lower than previous estimates for the Late Cretaceous and Paleocene and an order of magnitude lower than today.