Two species of oak are dominant in French forests: pedunculate oak (Quercus robur L.) and sessile oak (Quercus petraea Liebl.). Their differentiation is not straightforward but is essential to better understand their respective molecular content in order to better valorize them. Thus, to improve oak species identification, an untargeted UHPLC-HRMS/MS method associated with a two-step data treatment was developed to analyze a wide range of specialized metabolites enabling the comparison of both species of oak extracts. Pooled extracts from sessile and pedunculate oaks, composed of extracts from several trees of pure species from various origins, were compared using first the Venn diagram, as a quick way to get an initial idea of how close the extracts are, and then using a molecular network to visualize, on the one hand, the ions shared between the two species and, on the other hand, the compounds specific to one species. The molecular network showed that the two species shared common clusters mainly representative of tannins derivatives and that each species has specific molecules with similar fragmentation patterns, associated in specific clusters. This methodology was then applied to compare these two pooled extracts to unknown individuals in order to determine the species. The Venn diagram allowed for the quick presumption of the species of the individual and then the species could be assigned more precisely with the molecular network, at the level of specific clusters. This method, developed for the first time, has several interests. First, it makes it possible to discriminate the species and to correctly assign the species of unknown samples. Moreover, it gave an overview of the metabolite composition of each sample to better target oak tree utilization and valorization.