An alarming rate of ground water depletion and increasing labour scarcity are major threats to future rice production in north west India. Management strategies that reduce the irrigation amount and labour requirement while maintaining or increasing yield are urgently needed. Dry seeded rice (DSR) has been proposed as one means of achieving these objectives, but little is known about optimal water management for DSR. Therefore a field study was conducted on a clay loam soil in Punjab, India, during 2008 and 2009, to investigate the effects of irrigation management on the performance of puddled transplanted rice (PTR) and dry seeded rice. Irrigation scheduling treatments were based on soil water tension (SWT) ranging from ponding/saturation (daily irrigation) to alternate wetting and drying (AWD) with irrigation thresholds of 20, 40 and 70 kPa at 18–20 cm soil depth. Rainfall was above average and well distributed in 2008 (822 mm), and average and less well distributed in 2009 (663 mm). With daily irrigation, crop duration of PTR and DSR was the same. Shifting from daily irrigation to AWD delayed crop maturity by 8–17 d, and DSR was more affected than PTR, and more so in the drier year. Crop performance in terms of tiller density, leaf area index and growth rate was better in DSR than PTR with daily and 20 kPa irrigation scheduling. However, crop performance was poorer in DSR than PTR at higher (40 and 70 kPa) irrigation thresholds, more so in the drier year when DSR showed signs of severe iron deficiency which was not overcome with iron sprays. Yield components were similar in both establishment methods when irrigation was scheduled daily or at 20 kPa, but panicle density and the number of filled grains per panicle were significantly lower at 40 and 70 kPa in DSR than PTR. Each year, yield of DSR and PTR were similar when irrigation was scheduled daily or at 20 kPa. Yields of both PTR and DSR declined under higher water deficit stress (40 and 70 kPa irrigation thresholds), but more so in DSR, and more so in the drier year. There was a very large and significant decline in irrigation water input with irrigation at 20 kPa compared to daily irrigation in both establishment methods, but only a very small decline in irrigation amount when the threshold was increased from 20 to 40 and 70 kPa. Irrigation water use in DSR-AWD treatments was significantly lower than in respective PTR treatments (e.g. by 33–53% when irrigation was scheduled at 20 kPa). The results suggest the feasibility of reducing irrigation amount while maintaining yield by replacing PTR with DSR with AWD, provided that soil tension is kept lower than 20 kPa at 20 cm depth, but that this threshold needs to be tested over a wider range of seasonal and site conditions and varieties.