Permanent raised bed cultivation for rice-wheat cropping systems of South Asia is a paradigm shift from the conventional practice of planting on flat land. Conventional practice involves substantial tillage and puddling of soil, thereby destroying soil aggregates and promoting degradation of soil organic matter. Permanent raised bed cultivation employs a bed and furrow planting configuration that is permanently maintained for all crops with only periodic reshaping. Permanent beds offer the opportunity to reduce tillage and rebuild soil structure as well as soil organic matter. The bed and furrow configuration also reduces irrigation water inputs and improves nitrogen use. Three experiment station trials comparing conventional and permanent bed cultivation in a rice-wheat-mungbean rotation have been ongoing in Ranighat, Nepal; Nashipur, Bangladesh; and Rajshahi, Bangladesh since 2001. This paper will discuss trends in crop productivity, N response and differences in irrigation water inputs in these experiments. Permanent beds have generally outperformed conventional flat practice for all crops with the exception of wheat at the Ranighat site. Permanent beds improved rice and mungbean productivity more than wheat productivity. Mean yields were increased by 0.5-1.7 t/ha for rice and by about 0.2 t/ha for mungbean. Differences in mean wheat yields ranged from –0.2 to 0.5 t/ha. Rice and wheat yield trends on permanent beds over time at Ranighat and Rajshahi have either followed the variation observed with conventional practice or increased with time. However, at Nashipur, wheat yields on beds declined over time, which was not observed for conventional practice on the flat or with rice. Treatments to evaluate N response and placement (band or broadcast) effects were included in the experiments at Nashipur and Rajshahi. Mean yield response to nitrogen fertilization was generally greater on beds than on the flat and more with rice than wheat. Fertilizer placement has had little impact on yields. At the same level of N fertilizer input, nitrogen uptake in wheat, mungbean and rice grain at Nashipur and wheat and rice grain at Rajshahi were higher for the bed treatments than for the flat. At Rajshahi and Nashipur, inputs of irrigation water were reduced 21-38% in wheat, 26-33% in mungbean and 16-33% in rice. Consistent improvements in yields, improved use of fertilizer N and irrigation water, together with cost savings in labor/land preparation and seed inputs with permanent beds are convincing farmers in Bangladesh to adopt this innovative technology.