Nannofossil assemblages from the last 144 ka (∼Marine Isotope Stage, MIS6) were studied in a piston core (MD01-2421) from the NW Pacific (36°01.4′N, 141°46.8′E, WD: 2,224 m), off the coast of central Japan. The site is influenced by both the subtropical Kuroshio and the subarctic Oyashio Currents, and is located just north of the Kuroshio Front. This study attempted to reconstruct paleoceanographic changes in the region by illustrating the past migration patterns of the Kuroshio Current, which is considered to be one of the strongest western boundary currents in the Pacific. In order to reconstruct the past migration patterns of the Kuroshio Current, relative abundance values of several taxa classified as either “Kuroshio Species” or “Oyashio Species” were used as indices. Using these species, a ratio (Tn: Nannofossils temperature index), representing the mixing of the Kuroshio and Oyashio Currents, was calculated to express quantitatively the influence of the Kuroshio Current. Tn values fluctuated widely, ranging from 1.45 (135 ka) to 86.9 (115.3 ka), and corresponds to the warm–cold cycles obtained by oxygen isotope analysis. The fluctuation pattern of percentage abundance of Umbellosphaera spp. shows a strong similarity to that of the precessional cycle. It may indicate that subtropical water has influenced the Japanese coast during these warm intervals, possibly due to the expansion of subtropical divergence. Also, the fluctuation pattern of Tn shows a similar time-transgressive trend to that of the precessional cycle. This implies that latitudinal displacement of the Kuroshio Front was related to subtropical climatic variability. The ratio between Tn and Tp (planktonic foraminifera temperature index) drastically increased at about 20 ka, in MIS2, with small peaks in MIS6. This suggests that a thin layer of warm water may have covered the study area during these intervals. An increased preservation of organic matter (i.e. CaCO 3 preservation) possibly affected the absolute coccolith abundance (coccolith/g dry sediment), although other factors controlling absolute coccolith abundance (e.g. nutrient levels) need to be taken into consideration.