Soil Olsen P level has a major influence on crop yield,efficient P utilization,and soil fertility.In this study,the optimum Olsen P range was determined from long-term (1990-2012) field experiments in three typical soil types of China under single cropping of maize or double cropping of maize and wheat.The critical soil Olsen P value for crop yield was evaluated using three different models,and the relationships among P use efficiency (PUE),Olsen P.and total P were analyzed.The agronomic critical soil Olsen P values obtained from the three models for the neutral soil of Gongzhuling and the calcareous soil of Zhengzhou were similar;however,the values from the linear-linear and linear-plateau models for both maize and wheat were substantially lower than those from the Mitscherlich model for the acidic soil of Qiyang.The PUE response change rates (linear equation slopes) under different soil Olsen P levels were small,indicating slight or no changes in the PUE as the soil Olsen P increased in all three soils.A comparison of the Olsen P levels that achieved the maximal PUE with the agronomic critical values derived from the three models indicated that the linear-plateau model exhibited the best performance.The regression equation coefficients of Olsen P response to total P decreased as follows: Zhengzhou (73 mg g-1) ＞ Qiyang (65 mg g-1) ＞ Gongzhuling (55 mg g-1).The Olsen P level increased as the total P increased,which may result in a decrease in PUE.To achieve a relatively high crop yield,PUE,and soil fertility,the optimum Olsen P range should be 13-40,10-40,and 29-40 mg kg-1 at Gongzhuling,Zhengzhou,and Qiyang,respectively.