Optimum plant population and irrigation strategies for maize grown in the Dalby district of the Darling Downs in Queensland, Australia, were investigated using the APSIM crop simulation model. After testing the model against three seasons of experimental data, simulation experiments using different irrigation strategies were conducted across a range of plant populations ranging from 20 000 to 80 000 plants/ha, on two soil types with plant available water capacities (PAWC) of 146 mm and 220 mm. All soil type × plant population × irrigation strategy scenarios were simulated using the historical climate record for Dalby from 1889 to 2004, in order to obtain long-term average yield and gross margins (LGM) for each scenario. Soil water was reset to two-thirds of PAWC at sowing in each year. Plant populations required to achieve maximum LGMs ranged from 50 000 to 80 000 plants/ha across the range of scenarios, and were higher than currently recommended by district agronomists for partially irrigated maize. The use of higher plant populations increased season-to-season variability in grain yield and gross margins and may not be a suitable strategy for growers who do not want to increase their risk of crop failure. Partially irrigated maize achieved substantially higher gross margins in years where a positive Southern Oscillation Index phase was recorded in August, and the use of higher plant populations in such years also increased long-term profitability, but also increased the risk of crop failure. Economic gains were achieved by varying the timing and amount of irrigation within a limited available irrigation volume, with a single 100 mm irrigation giving greater LGMs than two 50 mm irrigation events on both soil types, when the irrigation events were scheduled to fill a soil water deficit equal to the effective irrigation volume. However, under full irrigation the use of smaller irrigation volumes increased LGMs on the 146 mm PAWC soil, demonstrating the importance of timely irrigation scheduling on low PAWC soils.