Concern about the impacts of two invasive annual brome grasses (cheatgrass and Japanese brome, Bromus tectorum L. and B. japonicus Thunb. ex Murray) on the mixed-grass prairie of North America's northern Great Plains (NGP) is growing. Cheatgrass is well known west of the NGP, where replacement of fire-intolerant, native sagebrush steppe by fire-prone, exotic annual grasslands is widespread. Consequently, fire is often not considered as a tool for controlling annual bromes. This should not be the case in the NGP, where mixed-grass prairie is adapted to frequent fires. Fire's efficacy may vary with the degree of invasion, though; suppressing postfire annual brome populations or enhancing the native plant community may improve postfire annual brome control in highly invaded areas. To test this, we performed an experiment at two sites to evaluate the relative effectiveness of prescribed fire alone, fire followed by imazapic application and fire followed by native seeding across a pretreatment invasion gradient of annual brome-to-native species cover ranging from 0.05 to 2.35. Fall-prescribed fire alone greatly reduced annual bromes, but by the second yr after treatment the effect was significant only at invasion ratios < 1.2. Postfire imazapic application reduced annual bromes even further than fire alone, but only for 1 yr at the less invaded site and only at invasion ratios > 1.2 in yr 2 at the other site. Native species cover and total species richness responded positively to all treatments, but the degree of their response varied along the invasion gradient, between sites, with time since treatment and among treatments. Also, at one site, fire yielded a lagged stimulation of short-lived, exotic forbs. Seeding had little effect. Fire is an effective tool for reducing annual bromes in the NGP at lower invasion levels, but more tools are needed for long-term, effective control at highly invaded sites.