The efficiency of variable rate seeding (VRS) was confirmed in various crops. It is proven that corn requires increasing seeding rates in high-yielding zones, whereas soybeans need lower rates. However, the data for wheat appeared to be controversial. The aim of our experiment was to determine the most efficient strategy for variable rate fertilization and seeding in spring wheat in the conditions of Canadian Prairies. Two approaches were tested: based on Normalize Difference Vegetation Index (NDVI), and on digital elevation model (DEM). The strategies for VRS included two types: seeding rate increased from 100 in the lowest yielding zone to 160 kg/ha-1 in the highest yielding area with the increments of 12 kg/ha-1, and maximized to 160 kg/ha-1 in the lowest yielding area with a gradual reduction to 100 kg/ha-1 in the highest yielding zone. The constant seeding rate was 130 kg/ha-1. The rates of carbamide varied from 93 to 284 kg/ha-1, the constant rate was 252 kg/ha-1.The experiment was laid out in four replications. Constant rate fertilization and seeding were used as control. The replications were placed within the field so that the average last year’s NDVI values in each variant were equal. At the peak of the growing season, NDVI was calculated for comparison of green biomass development in each variant. On average, NDVI in the variants with variable rate fertilization and seeding exceeded the constant rate by 9.2%. The difference in grain yield between the same variants was 6.7%. Both values were significant at P< .05. Variable rate seeding with the reduced rates in high-yielding areas was more efficient that the variants with the higher rates; however, the difference was not statistically significant. Also, no significant difference at P< .05 was found between the variants based on NDVI and DEM.