We estimated survival rates of 135 female greater sage-grouse (Centrocercus urophasianus) on 3 study areas in southeastern Oregon, USA during autumn and winter for 3 years. We used known-fate models in Program MARK to test for differences among study areas and years, investigate the potential influence of weather, and compute estimates of overwinter survival. We found no evidence for differences in survival rates among study areas, which was contrary to our original hypothesis. There also were no declines in survival rates during fall–winter, but survival rates varied among years and time within years. Average survival rate from October through February was 0.456 (SE = 0.062). The coefficient of variation for this estimate was 13.6% indicating good precision in our estimates of survival. We found strong evidence for an effect of weather (i.e., mean daily min. temp, extreme min. temp, snow depth) on bi-weekly survival rates of sage-grouse for 2 of the study areas in one year. Extremely low (<−15° C) temperatures over an 8-week period and accumulation of snow had a negative effect on survival rates during the winter of 1990–1991 on the 2 study areas at the higher (>1,500 m) elevations. In contrast, we found no evidence for an influence of weather on the low-elevation study area or during the winters of 1989–1990 and 1991–1992. Extreme weather during winter can cause lower survival of adult female sage-grouse, so managers should be aware of these potential effects and reduce harvest rates accordingly.