Climate model based projections of drought go back a very long way – to Charney in 1975 who identified the role of vegetation change in reinforcing and perpetuating rainfall droughts. Since them many publications have reported on how droughts will evolve in the future, but very few have discussed whether climate models can actually simulate the characteristics of drought. The connection between model evaluation, through model development to better predictions has been virtually absent and the feedback identified by Charney 45 years ago is still absent in many CMIP5/6 climate models. I will discuss the current state of drought simulation in climate models (getting better with some caveats), what land surface research is being implemented into climate models and what new research needs to be incorporated such that confidence in projections of the character of droughts can be improved.