The Cretaceous Earth, with its greenhouse climate and absence of major ice caps in the polar regions, represents an extreme scenario for modeling future warming. Despite considerable efforts, we are just at the verge of fully understanding the conditions of a warm Earth, and better, more extensive proxy evidence is needed to solve existing discrepancies between the applied temperature proxies. In particular, the Maastrichtian temperature trends are controversial, since data indicate cooling in the South Atlantic and contemporary warming of the North Atlantic. The “heat piracy” hypothesis involves northward heat transport to midlatitudes via oceanic currents and is used to explain the contrasting polar cooling/warming patterns. Here, we present Δ47 and δ18O data from nine coccolith-enriched samples from a shallow core taken from the Danish Basin (Chalk Sea), representing a key location at the northern mid-latitudes. Based on Δ47 data of coccolith-enriched material, sea-surface temperatures for the late Campanian–Maastrichtian ranged from 24 °C to 30 °C, with an average of 25.9 °C ± 2 °C. This is 4–6 °C higher than estimates based on Δ47 of bulk samples and 8–10 °C higher than reported temperatures based on bulk δ18O data from the same core. However, these higher temperature estimates are lower, but overall in line with estimates of Late Cretaceous tropical sea-surface temperatures from TEX86 (tetraether index of 86 carbons), when considering latitudinal differences. The study highlights the potential of clumped isotope paleothermometry on coccoliths as a valid, reliable proxy with which to reconstruct sea-surface temperatures.