Agricultural land in coastal zone of Bangladesh grows a low yielding single rice crop (aman) during the rainy season, due to lack of fresh water in dry season. To develop a double rice cropping system with an aus crop grown at the onset of the rainy season followed by an aman crop, a field experiment was conducted during 2006 – 2008 at Batiaghata in Khulna, to test the hypothesis that high system productivity can be achieved by: (i) dry seeding a short-duration aus variety and applying supplemental irrigation during establishment, and (iii) transplanting a high-yielding aman variety after harvest of the aus crop. Treatments in aus season included three water regimes, i.e., rainfed (I1); supplemental irrigation (I2), full irrigation (I3) and three sowing dates viz. early-April (D1), mid-April (D2), late-April (D3). In aman season, two high yielding varieties namely BR11 and BRRI dhan53 were evaluated. Yields with full and supplemental irrigation were comparable (4 t ha-1) and were greater than those of fully rainfed aus (2 t ha-1) when there were dry spells after sowing. Supplemental irrigation required much less irrigation water (100 – 200 mm) than full irrigation (660 – 1042 mm) but had similar yield. The average incremental irrigation water productivity varied from 0 to 1.3 kg grain ha-1 mm-1. Delayed planting of aman, decreased the growth duration of BRRI dhan53 and reducing yield compared to yield of earlier transplanting. The duration of BR11 was less affected by sowing date and had similar yield level (> 4 t ha-1) to that of BRRI dhan53 with early April sowing. Under favorable conditions, aus-aman cropping system yield ranged from 8.0 to t ha-1 yr-1.