Conventional tillage and crop establishment practices in the rice-wheat system require a large amount of water and labor, which are becoming scarce and costly day by day. A field experiment was conducted in the Indo-Gangetic Plains for five years (2002-2007) to evaluate various tillage and crop establishment practices for their efficiency in water and labor use and economic profitability. Yield of rice was higher with conventional puddled transplanting (7.2-8.8 Mg ha-1) than no-tillage direct seeding on flat land (6.5-7.3 Mg ha-1) and raised beds (3.1-6.3 Mg ha-1). Gap between the yields with puddled transplanting and no-tillage direct seeding widened to 2.0-2.8 Mg ha-1 after 2 yrs. Yields of wheat on flat land and raised bed following either puddling or no-tillage in rice were similar (4.0-5.3 Mg ha-1) and maintained over the yrs. In rice 2-25% less irrigation water was applied with no-tillage. With no-tillage and direct seeding of rice, there were less human labor (46-47 man d ha-1) and machine labor (7-8 tractor hr ha-1) use than with puddled transplanting (61-67 man d ha-1 and 14-15 hr ha-1, respectively). Conventionally-tilled wheat had a machine labor use of 12 tractor hr ha-1, whereas in other treatments it was 6 tractor hr ha-1. Rice on raised beds had the lowest economic return, which decreased over the yrs. Wheat in both raised beds and no-tillage, however, had higher returns than the conventional system because lower machine use. The study suggested that wheat can be successfully grown with no-tillage and on beds but the direct-seeding and bed planting technologies for rice need improvement on a site- and season-specific basis.