Maize may assume a more significant role in grain crop production systems in north-east Australia if the probability of producing low yields associated with given amounts of available water can be reduced. Growing hybrids with very early maturity provides a possible way to achieve this. Simulation studies of dryland maize production in areas of highly variable rainfall in north-east Australia were undertaken using long-term weather data input to the APSIM model configured for quick to medium maturity maize. The studies focussed on sowing time options, population density, cultivars, and water availability at sowing. Simulation outputs included predicted mean and median yield, measures of yield variability, and the probability of producing low to very low yield (<2 t/ha). The study showed that optimum sowing date varied with location, and that low populations gave more reliable production, despite some potential yield losses in favourable years. The results of the simulation study provide estimates of yield and thus economic viability of maize production that are interpreted in terms of seasonal variability. They indicate that maize is a viable dryland cropping option provided that cultivar, sowing time and starting water conditions are optimised. Non-optimal conditions of water supply at sowing should be avoided, as greater variability in yield and reduced viability are predicted.