Sedimentological, micropaleontological and geochemical studies of IODP Leg 302 boreholes, M0004A and M0002A, drilled on the Lomonosov Ridge near the North pole, are reported to construct chronostratigraphy and helped unravel Paleogene palaeoclimatic. These studies concluded to a mean annual temperature of sea surface waters ranging between 21 and 25 °C during the Paleocene–Eocene Thermal Maximum and to the occurrence of episodic ice on the Arctic shelf in the earliest Mid-Eocene. Pollen records are characterised by excellent preservation and identification of 112 taxa reveals a subtropical coastal vegetation comprising Taxodium-Glyptostrobus swamps alternating with Avicennia mangrove. The hinterland was occupied by a subtropical evergreen forest including Engelhardia and Distylium, occasionally replaced by Rhodoleia and Castanopsis-Lithocarpus. Climate was reconstructed using the Climatic Amplitude Method, resulting in the following temperature ranges for low elevations: mean annual temperature 18–22 °C, mean temperature of the coldest month 10–14.5 °C, mean temperature of the warmest month 26–27.5 °C, and annual precipitation 1150–1400 mm. Palaeovegetation data allow us to establish climatostratigraphic relationships with the generalized oxygen isotope curve, providing evidence for warm periods (e.g.: Paleocene–Eocene Thermal Maximum, PETM; Early Eocene Climatic Optimum, EECO; Mid Eocene Climatic Optimum, MECO) alternating with cooler periods. We date the latest record of Avicennia mangrove in the Arctic Basin to the MECO. The Azolla phase did not occur after the EECO but at ~50 Ma during a climatically unstable episode during the thermal optimum. Recurrence of rarely produced and weakly transported Avicennia pollen, the exceptional level of preservation of pollen grains, and the high plant diversity indicate that lands of the Lomonosov Ridge were connected to North America in the Ellesmere Island region which, in addition, provided pollen grains of Gymnosperms from high elevations. These physiographic features, suggesting the inflow of altitudinal cold waters into the sea, may solve the apparent contradiction between possible sea-ice and our reconstructed air temperatures during the Middle Eocene.