Climate is the principal driver of plant production across all rangeland ecosystems. Monthly or seasonal forecasts of rangeland plant production could provide valuable management information pertaining to livestock stocking rates, restoration planning, wildfire fuel loads, and wildlife management. In the sagebrush steppe ecoregion, plant production is dependent upon climate but also varies by geographic location and the ecological state or dominant plant community. In this study, we used retrospective seasonal climate forecasts and plant production models to determine the utility of monthly or seasonal forecasting (with up to 7-month lead times) across both geographical sites and ecological states. This study was conducted in 45 intact late-seral sagebrush steppe sites over a 10-year period across a 50,000 km² area in southeastern Oregon. Research sites were sorted into five geographical sites and into five ecological states. The objectives of this study were to (1) model plant functional group responses to actual and forecasted climate across time and space; (2) determine plant production forecast skill for lead times of up to 7 months; and (3) differentiate plant production model and forecast skill among both geographical sites and ecological states. We found that plant production models and forecasts were able to identify key functional group responses to precipitation and temperature inputs. We also determined that climate and plant production forecasts could be used to produce significant monthly plant production forecasts for lead-times of up to 7-months for all site, state, and plant functional groupings. We also determined that forecasts of total plant production by ecological state were significantly better than forecasts based on geographical sites. Forecasting by ecological state could lead to better rangeland management decisions as it matches management inputs with inherent site conditions.