Context The Acland Land System overlying the Walloon sandstone coal deposits in southern Queensland is generally marginal for cropping but well suited to grazing, and thus cultivated land is commonly returned to pasture. Rehabilitation of these lands after open-cut coal mining seeks to be safe, stable and self-sustaining to satisfy requirements for ecologically sustainable development. Aims The present paper evaluates the sustainability and economic viability of beef production on (a) lands retired from cultivation and then rehabilitated with sown pastures after open-cut coal mining at the New Acland mine site, and (b) similar nearby pasture lands that were not mined but were also retired from cultivation. Methods The GRASP grazing systems model was modified and calibrated with short-term (5-year) grazing trial data (soil, pasture and cattle observations), and then used with long-term (60-year) weather data to estimate effects of land type, pasture rundown, climate and grazing pressure on productivity and economic returns. The productivity of three rehabilitated sites and 15 unmined sites were evaluated, including pastures on six commercial properties. Key results Estimates of long-term mean annual growth of pastures on unmined lands retired from cultivation on three land types (Mountain Coolibah, Brigalow Uplands and Poplar Box) were 3398, 2817 and 2325 kg/ha respectively. Pasture growth was greater on rehabilitated lands; 3736 kg/ha on the site most typical of rehabilitated lands and a mean of 4959 kg/ha across three sites. Seasonal conditions had large effects on cattle liveweight gain (133–213 kg/head per year during the trial); however, pasture growth was the main driver of beef production and economic returns per hectare. In GRASP, potential nitrogen uptake was used to influence key pasture growth processes and accounted for 64% of variation in observed annual growth. The short-term lift and subsequent rundown in productivity typically associated with sown pastures was estimated to have increased mean annual pasture and cattle productivity during the 2014–2018 trial period by up to 17% and 25% respectively. Estimates of long-term mean annual beef production and economic returns for the unmined lands were less than estimated for rehabilitated lands and were 139 kg/head.year (45 kg/ha.year) and AU$154/adult equivalent. Conclusions Rehabilitated lands were found to be sustainable for beef production at grazing pressures up to 30% utilisation of annual pasture growth, and comparable with grazing systems on native and sown pastures in good condition. Pastures on unmined lands retired from cultivation had reduced productivity. Implications Overgrazing is a significant and on-going residual risk to sustainable production. Grazing regimes need to continually adjust for changes in novel landscapes, pasture condition and climate. The methods used in the present study could be applied more generally.