High mountain environments are among the most fragile on Earth. Due to anthropogenic disturbances and the exposure to extreme weather events, the rates of soil erosion have recently been accelerating, resulting in ecological degradation and geological hazards. Ecological restoration of mountains and an improved understanding of nature-based solutions to mitigate land degradation is therefore of utmost urgency. Identifying hotspots of soil erosion is a first step towards improving mitigation strategies. A promising methodology to identify erosion hotspots is sediment source fingerprinting, that differentiates the properties of soil from different sources, using signatures such as elemental geochemistry or radionuclides. However, in areas with complex lithologies or shallow and poorly developed soils, geochemical fingerprints allow only a rough distinction between erosion hotspots. In this opinion paper, we explore the relevance of environmental DNA (eDNA) that originates from plant litter and fixes onto fine soil particles, as a targeted sediment fingerprinting method sensitive to vegetation that could potentially allow the identification of erosion hotspots and their relative importance from sedimentary deposits. Pioneering studies indicate that eDNA allows not only the detection of specific vegetation communities, but also the identification of individual plant species. Supported by the increasing availability and quality of vegetation maps and eDNA reference libraries, we argue that sediment source fingerprinting using eDNA from plant litter, will evolve into a valuable method to identify hotspots of soil erosion and allow stakeholders to prioritize areas where ecological restoration is necessary in high mountain environments.