[Objectives]  DNA barcoding has been widely applied to many insect taxa but not a large number of the Lepidoptera, especially smaller and medium-sized moths, which are difficult to identify morphologically. It is not yet possible to build a complete DNA barcoding system. This study aims to provide data sources and support for the construction and improvement of DNA barcoding systems for lepidopteran pests, verify the accuracy of the COⅠgene as a DNA barcoding gene, explore the feasibility of using the 28S D2 gene as a DNA barcoding assistant gene, and test the success rate of BOLD system identification. [Methods]  Morphological identification and DNA sequencing of 490 small and medium-sized moths collected in the Baiyanggou Scenic Area in Beijing were carried out. Intraspecific and interspecific distances were then calculated based on differences in the COⅠand 28S D2 genes and NJ phylogenetic trees were constructed. [Results]  The BOLD system had an average success rate of 65% but the success rate for smaller moths and noctuids was low. NJ trees based on the COⅠand the 28S D2 genes successfully identified 94.4% and 89.4% of specimens, respectively. [Conclusion]  The COⅠgene is suitable for general DNA barcoding of moths whereas the 28S D2 gene is more conservative and therefore not suitable as an assistant gene for DNA barcoding.