At present, the research on remote sensing inversion of heavy metal in water is still relatively weak, therefore the study on the spectral characteristic of heavy metal polluted water in nature is an important basic work, which is an important theoretical basis for the band selection when realizing the remote sensing inversion, and the measuring results are also the important parameters necessary for the remote sensing inversion model in the future. Firstly, using Analytical Spectral Devices (ASD) spectrometer, measuring the water-leaving reflectance spectrum curve of mine drainage of Dabaoshan Mountain as an example of typical heavy metal polluted water under two different water depths and light conditions, we found that there was a stable reflection peak in 600~700 nm (red light). Then further comparing the reflection peak position of mine drainage of Dabaoshan Mountain with two types of water (turbid water and eutrophication of water) which are common in nature, we found that the reflection peak position of Changhu Reservoir near the Quartz Factory as an example of turbid water was in 550~700 nm (green and red band) and that of Beijiang River near the Shaoguan Smelter as an example of eutrophication of water was in 550~600 nm (green band), and the position of the reflection peak of these three kinds of water was different, which means that the reflection spectrum of this heavy metal polluted water has a good separability with these two common types of water. On this basis, through a combination of water quality remote sensing model and the indoor water extinction coefficient measurements, we obtained the scattering coefficient and absorption coefficient spectrum of the mine drainage of Dabaoshan Mountain, and further eliminated water molecules absorb effect, finally got the absorption spectrum curve of compositions in this heavy metal polluted water, with the results showing that: it absorbed the strongest in purple band, while the weakest in red band; Starting from 400nm, the absorption coefficient decreased rapidly, then slowdown in the blue and green light band; arrived at the yellow light band, it decreased rapidly again until 676 nm, which reached the minimum; then it increased rapidly to 750nm, and then the change slowed down. Finally, combined with the water quality test results of water samples, the causes of the spectrum of the heavy metal polluted water were analyzed, and we found that the water color in-situ and its absorption coefficient spectrum characteristics were consistent with the color and absorption coefficient of ferric sulfate solution measured in our previous study, therefore we considered that the spectral characteristic of this water sample was caused by the ferric sulfate and its hydrolysate. The results above showed that the absorption spectrum of the mine drainage of Dabaoshan Mountain had obvious characteristics, and the position of the reflection peak and strongest absorption wavelength was clear, which were the important characteristic bands for future extraction of heavy metal concentration in water using satellite remote sensing technique. In this paper, the reflectance spectrum, extinction coefficient spectrum, scattering coefficient spectrum and absorption coefficient spectrum of the mine drainage of Dabaoshan Mountain as an example of typical heavy metal polluted water were obtained for the first time, which provides a method basis for the optical parameters inversion of drainage in other heavy metal mines and also lays a good theoretical foundation for the quantitative extraction of heavy metal concentration in water using remote sensing technology in the future.