Dumasia taxonomy and classification have long been problematic. Species within this genus have few morphological differences and plants without flowers or fruits are difficult to accurately identify. In this study, we evaluated the ability of six DNA barcoding sequences, one nuclear (ITS) and five chloroplast regions (trnH-psbA, matK, rbcL, trnL-trnF, psbB-psbF), to efficiently identify Dumasia species. Most single markers or their combinations identify obvious barcoding gaps between intraspecific and interspecific genetic variation. Most combined analyses including ITS showed good species resolution and identifi-cation efficiency. We therefore suggest that ITS alone or a combination of ITS with any cpDNA marker are most suitable for DNA barcoding of Dumasia. The phylogenetic analyses clearly indicated that Dumasia yunnanensis is not monophyletic and is separated as two independent branches, which may result from cryptic differentiation. Our results demonstrate that molecular data can deepen the comprehension of taxonomy of Dumasia and provide an efficient approach for identification of the species.