Abstract. Cattle grazing is a potential post-mining land-use option for open-cut coal mines in the dry sub-tropical region of central Queensland, Australia, but no research has been conducted to determine the grazing capacity of these lands. A study was conducted to develop a model for estimating pasture productivity of rehabilitated mined lands, from which long-term sustainable stocking rates could be predicted. Rainfall-use efficiency (RUE), a reliable indicator of pasture productivity in this moisture-limited environment, was calculated for 17 plots across three minesites over a single growing season, and related by linear regression and stepwise multiple linear regression to several site and mine-soil properties. Plots were dominated by Cenchrus ciliaris (buffel grass), and ranged in age from 3 to 25 years since establishment. Slope (r 2 =0.45) and surface cover (r 2 =0.44) were most strongly correlated with RUE. These factors were interpreted as affecting surface retention of rainfall. The factors most correlated with RUE from multiple linear regression were slope (r 2 = 0.45), surface soil exchangeable Mg (cumulative r 2 = 0.71) and surface exchangeable sodium percentage (ESP) (cumulative r 2 = 0.77). ESP is a measure of soil dispersion and surface crusting, which when combined with slope (negative correlation), influenced the ability of incident rainfall to enter the soil profile. Mg was interpreted as a surrogate soil fertility factor, as Mg was strongly correlated with soil total N (r 2 =0.53) and cation exchange capacity (r 2 =0.74). Dry matter yield and RUE results are generally consistent with those observed on unmined pastoral lands in the region, but data from additional sites and over more seasons are required to fully develop and validate the model for minesite conditions.