Past climate oscillations have led to the presence or absence of sea ice in the North Pacific Ocean, which in turn has driven changes in community compositions. Marine diatoms (Bacillariophyceae) are unicellular, siliceous organisms that can respond sensitively to environmental change such as sea surface temperatures. Moreover, they are the dominant group of sympagic microalgae in the Arctic with a few species producing the sea ice proxy IP25 – a highly branched isoprenoid with 25 carbon atoms. So far, marine diatom records have been mainly analysed by microscopic observations, while only a handful of studies attempted a paleogenetic diagnostic. Here, we present an 18,000 year record based on ancient sedimentary DNA metabarcoding of diatoms for the Northwest Pacific Ocean and tackle the question if this approach is suitable as a new proxy for sea ice reconstructions. Our study site is located near the coast of Kronotsky Peninsula (Kamchatka). We collected 54 marine sediment samples from the 9.05 m long core SO201-2-12KL, which was taken at 2173 m water depth (N 53.992660, E 162.375830) and covers the time period since the deglaciation after the Last Glacial Maximum. In this core IP25 was previously detected in time slices representing Heinrich event 1 (15.1 ka BP) and Younger Dryas (12.2 ka BP). Since chloroplast DNA is present in several copy numbers in diatoms the probability of its long-term preservation in marine sediments is very likely. Therefore, we applied a short metabarcode of the chloroplast rbcL gene (76 bp), which was amplified for each sample three times independently. Subsequent PCR products were prepared for high-throughput sequencing. In a pre-study with a subset of only 12 sediment core samples all four classes of diatoms were detected, expectably centric diatom genera like Porosira, Chaetoceros and Thalassiosira dominated the dataset. Genera containing ice-associated diatoms, such as Melosira, Rhizosolenia, Navicula or Haslea were recovered, but sparsely. Based on our preliminary results, we await high quality data for the larger data set. By now, our study demonstrated that DNA metabarcoding of marine diatoms has the potential to support and complement the analysis of long-term developments of community composition and diversity changes, e.g. in areas of low diatom preservation due to silica dissolution.