Rice (Oryza sativa L.) cultivation in Sri Lanka is restricted by limited water availability. Increased variability of seasonal rainfall as a result of climate change will further exacerbate risk in rice cultivation. Knowing the onset of rainfall through forecasts, adjusting planting date based on the predicted rainfall onset, and efficient irrigation management are important determinants to reduce the risk in rice farming. In this study, a modelling approach was used to assess the yield advantage of changing planting dates according to the onset of seasonal rainfall compared to a fixed planting date in the major rainy season (Maha), and evaluated the efficiency of irrigation management. The model, Agricultural Production Systems Simulator (APSIM), was parameterized using data from the literature, and the model could simulate the observed yield of rice varieties Bg300 and Bg359 with a good fit (CV<14%). Simulations carried out for Bg300, the rice variety that has recorded the largest extent of cultivation in Sri Lanka, revealed that changing the planting date according to the onset of rainfall can reduce the irrigation water requirement and yield loss. Early onset coupled with early planting resulted in an average yield of 3900 ± 150 kg/ha, which was 33 % higher than that of late planting with the same onset. Late onset coupled with late planting resulted in an average yield of (3930 ± 200 kg/ha), which was 34% higher than that of early planting with a late onset. Moreover, late planting at a late onset resulted in a higher variability in water productivity (4.3 ± 0.34 kg/ha/mm) when compared to an early planting at an early onset (4.4 ± 0.12 kg/ha/mm), even though the mean water productivity was similar. Therefore, timing of rice planting should be adjusted based on the forecast of the onset of seasonal rainfall to harness the maximum potential of available natural resources.