Reliable establishment of pasture systems with a desirable composition of grasses and legumes remains a challenge in the highly variable climates of the northern grain belt of southern inland Queensland. In this paper, an alternative system is proposed whereby legumes and grasses are planted simultaneously in alternative strips. A 4-year study tested the establishment, production, and botanical composition of the tropical legume species Macroptilium bracteatum cvv. Cardaga and Juanita, Lablab purpureus cv. Endurance, and Clitoria ternatea cv. Milgarra sown as pure stands or with grass strips. The grass strips, and a grass-only treatment, both contained a mix of Panicum maximum cv. Petrie, Dichanthium aristatum cv. Floren, and Bothriochloa insculpta cv. Bisset. L. purpureus was relatively unproductive, yielding ~2000 kg/ha dry matter (DM) produced in each of the first 2 years. M. bracteatum produced 2050 and 3300 kg/ha DM in Years 1 and 2, with declining plant populations and DM in subsequent years. C. ternatea produced 960 kg/ha DM in Year 1, 2730 kg/ha DM in Year 2, and continued to persist throughout the trial, albeit at low DM production levels. The grass-only treatment was dominated by the sown grass species, while the legume-based treatments were dominated by a colonising native species, Dichanthium sericeum, in the fourth and final year. Sowing grass strips adjacent to the legume areas proved a successful strategy, with the proportion of sown grasses in the legume strips increasing to >20% of total DM by Year 4. Soil carbon changes did not differ between treatments, but total C in the top 0.15 m increased from 0.99% in Year 1 to 1.13% in Year 4, representing a net gain of 6.5 t/ha in C over 3 years.