A field experiment to evaluate the effect of N, K and time of K application on the growth, yield and chemical composition of wheat (Triticum aestivum) grown in saline soil with subsurface drainage, was conducted at the Central Soil Salinity Research Institute, Karnal Research Farm, Sampla during the winter seasons of 1992–93 to 1994–95. The treatments consisted of three levels each of N (0, 120 and 150 kg ha−1) and K (0, 50 and 75 kg ha−1) with the K applied at two different times (full basal and half basal + half top dressed 30 d after sowing). The growth characteristics (plant height, number of tillers and dry matter weight m−2 recorded at 60 d after sowing) and yield-attributing parameters (number of productive tillers and length of spikes) increased significantly with increasing N levels up to 150 kg ha−1 and K levels up to 50 kg ha−1. Similar responses of N and K were also observed on grain and straw yields of wheat. Time of K application had no significant effect on growth and yields. The N concentration in grain and straw increased significantly with the application of 120 kg N ha−1 over control. Application of K had no significant effect on the concentration of N in grain and straw. The K concentration in grain and straw increased significantly due to the application of 50 kg K ha−1 but it was unaffected by the time of K application. Application of K increased the efficiency of utilization of applied N. A trend towards declining salinity of the soil profile due to leaching through the subsurface drainage system was observed from 1992–93 to 1994–95. For three consecutive years the status of available N and K in surface soil (0–30 cm depth) increased due to their respective applications. This field study indicates that 150 kg N and 50 kg K ha−1 should be applied under canal-irrigated conditions to get the sustainable and optimum yield of wheat in saline soils.