Most soils used for agriculture in south-western Australia are sandy and are now deficient in both potassium (K) and nitrogen (N) for cereal and canola (oilseed rape; Brassica napus L.) grain production. However, the effect of applying different levels of both fertiliser K and N on grain yields of these crops is not known. We report results of 10 field experiments, conducted on sandy soils in the region, to measure the effects of applying both K and N on canola grain yields and concentration of oil and protein in grain. Four levels of K (0–60 kg K/ha as potassium chloride) and four levels of N (0–138 kg N/ha as urea) were applied. Significant grain yield responses to applied N occurred in all experiments for the nil-K treatment and each level of K applied, with responses increasing as more N was applied. For all levels of N applied, significant grain yield responses occurred when up to 30 kg K/ha was applied, with no further significant grain yield responses occurring when 60 kg K/ha was applied. The K × N interaction was always significant for grain production. Application of K had no effect on the concentration of oil and protein in grain. Application of N consistently decreased concentration of oil and increased concentration of protein in grain. The K × N interaction was not significant for concentration of oil or protein in grain, but application of up to 30 kg K/ha significantly increased canola grain and so oil yields (concentration of oil in grain multiplied by grain yield). Our results are likely to be relevant for all acidic to neutral sandy soils worldwide used for growing canola crops.