Research to study the effect of cover crop sowing rate, row spacing and cultivar on establishment of lucerne (Medicago sativa L.)-based pastures in a semi-arid zone mixed-farming system was undertaken at three sites over three consecutive years in south-eastern Australia. The weather of these 3 years subjected the establishing swards to a vastly differing range of moisture conditions. Conditions were so dry in the first year that lucerne swards failed to establish although it appeared at an early stage that there were advantages when lucerne was sown as a pure sward. By contrast, in the 2nd and 3rd years, moisture conditions at two of the three sites were so favourable that satisfactory densities were attained in all treatments. Lucerne forage yield was a more sensitive parameter than density to establishment method as it illustrated trial treatment effects during these wetter years when lucerne density alone could not. Thus, greater lucerne forage yields in the first summer after cover crop harvest under pure-sward establishment in comparison with establishment under a cover crop were seen at both Ardlethan and Barellan over both the 2nd and 3rd years of trials. Furthermore, the trial at Goolgowi in year 3 showed that when moisture was moderately limiting lucerne establishment, and yield is benefitted by sowing under a short, erect barley cultivar than under a taller and more spreading barley variety. This research highlights the greater lucerne forage production attained when lucerne is sown either as a pure sward or under cover crops with low levels of biomass.