Organic-walled dinoflagellate cysts and acritarchs are a vital tool for reconstructing past environmental change, in particular in the Neogene of the high northern latitudes where marine deposits are virtually barren of traditionally used calcareous microfossils. Yet only little is known about the paleoenvironmental value of fossil assemblages that do not have modern analogues, so that reconstructions remain qualitative. Thus, extracting their paleoecological signals still poses a major challenge, in particular on pre-Quaternary timescales. Here we unravel the relationship between species relative abundance and sea surface temperature for extinct dinoflagellate cyst and acritarch taxa from the Neogene of the Iceland Sea using palynological assemblages and organic geochemical (alkenone) data generated from the same set of samples. The reconstructed temperatures for the Miocene to Pliocene sequence of Ocean Drilling Program Site 907 range from 3 to 26 °C and our database consists of 68 dinoflagellate cyst and acritarch samples calibrated to alkenone data. The temperature range of five extant species co-occurring in the fossil assemblage agrees well with their present-day distribution providing confidence to inferred temperature ranges for extinct taxa. The 14 extinct dinoflagellate cyst and acritarch species clearly exhibit a temperature dependency in their occurrence throughout the analysed section. The dinoflagellate cyst species Batiacasphaera hirsuta, Labyrinthodinium truncatum, Cerebrocysta irregulare, Cordosphaeridium minimum, Impagidinium elongatum and Operculodinium centrocarpum s.s., and the acritarch Lavradosphaera elongatum, which are confined to the Miocene, have highest relative abundances and restricted temperature ranges at the warm end of the reconstructed temperature spectrum. The latter five species disappear when Iceland Sea surface temperatures permanently drop below 20 °C, thus indicating a distinct threshold on their occurrence. In contrast, species occurring in both the Miocene and Pliocene interval (Batiacasphaera micropapillata, Habibacysta tectata, Reticulatosphaera actinocoronata, Cymatiosphaera? invaginata) show a broader temperature range and a tolerance towards cooler conditions. Operculodinium? eirikianum may have a lower limit on its occurrence at around 10 °C. The calibration of species relative abundance versus reconstructed sea surface temperature provides a quantitative assessment of temperature ranges for extinct Miocene to Pliocene species indicating that temperature is a decisive ecological factor for regional extinctions that may explain the frequently observed asynchronous highest occurrences across different ocean basins. It demonstrates that qualitative assessments of ecological preferences solely based on (paleo) biogeographic distribution should be treated with caution. In addition to enhancing knowledge on marine palynomorph paleoecology, this study ultimately improves the application of palynomorphs for paleoenvironmental reconstructions in the Neogene of the Arctic and subarctic seas, a region essential for understanding past global climate.