A broad palette of natural dyes is often mentioned with reference to dyed textiles from ancient China. However, few scientific works address the problem of correctly identifying these dyes, often referring simply to unidentified sources.The aim of this work was the creation of a database of mass spectra of molecules related to Chinese natural dyes, intended to be used for their identification in archaeological textiles. High pressure liquid chromatography coupled with electrospray ionisation and quadrupole time-of-flight detection (HPLC-ESI-Q-ToF) was used to analyse twenty-eight reference specimens of Asian dyes, including relatively common dyes, such as sappanwood (Biancaea sappan), gromwell (Lithospermum erythrorhyzon), safflower (Carthamus tinctorius), madder (Rubia tinctorum, Rubia cordifolia and Rubia akane), and unusual dyes, such as rosewood (Dalbergia sp.), rhubarb (Rheum emodi), dragon's blood (Daemonorops sp.), gamboge (Garcinia hanburyi), violet (Viola yedoensis), etc. All analyses were repeated after artificial ageing, thus showing which molecules are more likely to be considered as markers for the identification of these dyes in ancient textiles. Because of the dye sources chosen, most molecular classes of natural organic colourants were included in the mass spectra database, i.e. anthraquinones, naphthoquinones, flavonoids, neoflavonoids, alkaloids, chalcones, carotenoids, xanthenes, polyphenols, etc. The information obtained from the reference specimens was then used to identify the dyes in thirty-one textiles from Dunhuang (northwestern Gansu, China, 7th-10th century AD) in the British Museum's collection. The sampling was guided by the data previously gathered from multispectral imaging and fibre optic reflectance spectroscopy (FORS), presented in a separate paper. Mixtures of dyes were commonly found, leading to hypothesise the re-dyeing or recycling of some of the textiles. A very broad palette of dyes suggested possible different origins for these textiles and a tendency to experiment with dyeing. These data, integrated with the ones obtained non-invasively, represent a significant source of information for researchers investigating textiles from the Silk Road.