Grasslands are extensive, surprisingly biodiverse, highly altered by humans, and not as well protected as other biomes. Restoration provides an opportunity to reverse degradation and increase local biodiversity. Here, I review emerging issues that will become increasingly important to the science and practice of restoration ecology. First, the global change dilemma. Restorations typically target species that were dominant before the Industrial Revolution, in effect, looking back in time. However, increasing atmospheric CO2 and methane, temperature, and nutrients, which are already having significant effects, will result in novel conditions that are unlike the past. Biotic introductions have occurred concurrently with climate change, altering the seed bank and propagule pressure from surroundings. Designing seed mixes with high diversity will increase the likelihood that species will be present that respond favorably to changes. Second, more research is needed on persistence as a longterm measure of stability. What is perhaps most important to restoration is how persistent restorations are on decade to century scales, and restorations are now of sufficient quantity and age to test questions about persistence. Third, the importance of stochastic processes due to priority effects have been supported by recent studies and have challenged the deterministic assembly model. Target species establishment could be improved by changing the order of introduction of species. Finally, grasslands provide many ecosystem services to society, including nutrient capture, food production, carbon storage, tourism and recreation, and nectar and pollen production. Grasslands are important culturally as outdoor science laboratories. For these reasons, I suggest that grasslands provide an excellent model system for restoration ecology.