Understanding the winter diets of endangered ungulates is critical for developing effective conservation strategies in temperate ecosystems, where seasonal fluctuations significantly affect food availability. This dataset contains DNA metabarcoding data used to compare the winter diet of the Korean long-tailed goral (Naemorhedus caudatus) between natural foraging and post-wildfire supplementary feeding. A total of 36 fecal samples were collected between January and February 2024 from five regions in South Korea (Hwacheon, Chuncheon, Donghae, Goseong, and Uljin). The dataset includes paired-end sequence reads in FASTQ format generated using Illumina MiSeq, along with metadata describing sampling site. Each file represents a single fecal sample, containing plant-derived DNA sequences amplified with universal chloroplast trnL primers. The data reveal region-specific dietary compositions dominated by woody taxa within Rosales, particularly Morus, Prunus, and Rosa, with Morus accounting for over 85% of reads in Uljin where supplementary feeding followed a large wildfire in 2022. The dataset provides baseline information for comparative dietary ecology, habitat management, and winter feeding strategies of N. caudatus, and can be reused to explore seasonal or interspecific variation. Samples were collected non-invasively.