To help farmers in the mid hills of Nepal improve their crop management and rice yields, we conducted a study testing different nursery management options and their effect on grain yield and yield components under rainfed conditions. The experiment was conducted in a farmer’s field during the cropping season 2009 and 2010 at Sundarbazar, Lamjung, Nepal, using a 3-factor factorial RCB design with 3 replications. The three management factors tested were 1) fertilizer management in the nursery, 2) seeding density, and 3) seedling age at transplanting, using the rainfed lowland rice variety Radha-4. There were eight treatment combinations, consisting of two levels of fertilization (0 and 20:20:0:13 kg NPKS ha^-1 at 15 DAS), two levels of seeding density (607 and 303 g·m^-2) and two seedling ages (20 and 40 days old). Two years’ results showed that top-dressed fertilizer in the nursery had no consistent effect on grain yield. However, lower seeding density (303 g·m^-2) resulted in taller plants, more productive tillers m^-2, less sterility and higher grain yield. In addition, older seedlings (40 days) produced taller plants, more productive tillers, more filled grains, and a higher grain and straw yield. The interaction analysis between both factors indicated that 40 days old seedling with a low seeding density produced the highest grain yield, both in the drought season 2009 and the high-yielding season 2010. The economic analysis confirmed that the treatment with low seeding density and 40 days old seedlings produced by far the highest net returns and B:C ratio in both seasons, independent of the fertilizer treatment. The combination of these two management components is therefore economically viable and profitable, and can be recommended to farmers. However, the results need to be confirmed for other varieties used by farmers in the region.