Green water is vital to vegetation recovery in karst area. Considering the processes in green water cycle, this paper coupled the canopy interception process, soil moisture movement process and evaportranspiration process, and built a green water cycle processes model to simulate and analyze green water cycle of typical vegetation types in karst area. Under the rainfall simulation experiment calibration and soil moisture monitor calibration, the model was extended to regional scale, and was used to simulate the green water cycle in a karst area, Guizhou Province, China, during October 2005 to March 2006. The results show that: First, most of the rainfall during this period turns into green water through canopy interception and infiltration into soil, which totally occupies 87.4% of rainfall and is supplied for vegetation ecosystem water use. Second, there are many differences in green water cycle processes of different vegetation types in the study area. The percentages of green water of different vegetation types are 93.3%, 93.2%, 91.5%, 81.9% for shrubs, coniferous woodland, mix woodland, grassland respectively. It can be concluded that with the vegetation recovery from grassland to shrubs and woodland in the study area, there will be more and more rainfall turning into green water which is used by the vegetation ecosystem and benefits the vegetation recovery. Third, the changes of green water cycle between months in different vegetation types share the same characteristics. The green water storage is increasing in October, January, and March in all vegetation types, and is decreasing in November, December, and February. In this period, the amount of green water is much greater than that of blue water, and the green water storage is increasing overall. Drought is unlikely to occur during October to March and the green water storage is supplemented in this period before the "spring drought" and "summer drought" in this karst area.