Abstract: The paper discussed the application of 3414 field trial design in soil testing and fertilizer recommendation using three years data collected in Hailaer, Inner Mongolia. The results showed phosphorus fertilizer recommended rate were 120, 90, 60 and 0 kg/ha when the soil available P (Olsen-P) was classified as low, medium, optimum and above optimum (Olsen-P 7 mg/kg, 726 mg/kg, 2650 mg/kg and 50 mg/kg) based on the relative yield and the model of quadratic equation in three variables. Soil available K were at relative high levels in most of the soil but potassium fertilizer was still recommended because of the dry climate condition in this region. The recommended potassium fertilizer rate was 2060 kg/ha as a risk fertilizer management strategy. Considering fertility field trials is an expensive and time consuming task and the results from field trial were incomplete or incorrect sometimes, it was emphasized the field trials for fertilizer rate should not carried out unless the soil testing results are available. A fixed rate and ratio of NPK in 3414field trial for all soil nutrient condition should not be adopted for one region. The type of mathematical model selected on the identification of the nutrient rate required for optimum yield is not only meaningful and significant trends (P 0.05) in mathematics but also practical and useful in agricultural production and agricultural chemistry philosophy, which fit the law of diminishing returns of crop response to fertilizer.