With wheat yields below normal for two consecutive years (1995 and 1996) in most Oklahoma wheat fields due to abnormal weather conditions and disease pressure, soil nutrients probably had accumulated at a level that would allow application of lesser amounts of fertilizers to produce normal yields the following year. A free wheat soil testing and education program was initiated to promote statewide soil testing for improved fertilizer recommendations and for helping farmers to cut wheat production cost. This was offered to wheat producers from June 15 to August 15, 1996 by Oklahoma Cooperative Extension Services. Three thousand and seventy‐nine surface (0–6 inches) and 2,957 subsurface (6–24 inches) soil samples were sent to Oklahoma State University (OSU) Soil, Water and Forage Analytical Laboratory over a two‐month period. Surface soil samples were analyzed for pH, buffer index (BI) if pH was less than 6.5, nitrate‐nitrogen (NO3‐N), available phosphorus (P) index, and available potassium (K) index. Subsurface soil samples were analyzed for NO3‐N only. Twenty‐four informational meetings, with a total attendance of 980, were conducted at the end of the program to help fanners interpret soil testing results and understand fertilizer recommendations. Topics at the meeting also included fertilizer nutrient management, soil‐plant‐nutrient interactions, economics of liming low pH soils. Soil testing results showed that 39% of the wheat fields had soil pH less than 5.5. About 50% and 84% of the fields did not need any P and K, respectively. Significant amounts of NO3‐N were found in subsoil samples. By combining NO3‐N from the surface and subsurface samples, there were 54% of the fields that had NO3‐N greater than 80 lb A‐1which is enough nitrogen (N) for a yield goal of 40 bu A‐1. Farmers would save more money on fertilizer cost if the fertilizer programs had been based on recommendations from soil testing reports. This program clearly demonstrated the need for regular soil testing and the importance of taking subsoil samples for estimating residual N.