How the suitable distribution of a given grassland type will respond to policy implementation and climate change is a prerequisite and the basis for effective grassland management and sustainable development. The maximum entropy (MaxEnt) model as a comprehensive evaluation method at species level was successfully used to simulate the temporal and spatial changes in the geographical distribution of Kobresia meadow in China’s Three-River Headwaters (TRH) region under the context of policy implementation and climate change. There were three main findings. First, there are the six dominant environmental variables affecting the distribution of Kobresia meadow (these being elevation, precipitation variation coefficient, March precipitation, April precipitation, April minimum temperatures, and May minimum temperatures: their threshold values are 3500–4000 m, 86–100, 8.5 mm, 12 mm, –3.5°C and 1.5°C, respectively, which indicate the most suitable conditions for Kobresia meadow distribution). Second, under current (2005–2014) climate conditions, the suitable distribution of Kobresia meadow gradually decreases from east to west, with the largest area in the east. Implementation of the ecological conservation policy in the TRH has greatly improved the ecological environment in lower elevation and high population areas. This has benefited the growth of Kobresia meadow and enlarged its suitable areas. Third, compared with the current distribution, areas suitable for Kobresia meadow will increase by 25.4–33.0% in the 2050s (2041–2060) and 17.4–33.0% in the 2070s (2061–2080) under four Representative Concentration Pathway (RCP) scenarios. More importantly, from the 2050s to the 2070s, areas suitable for Kobresia meadow will decrease or stabilise, bringing new challenges for protection and utilisation.