A field experiment with three-factor (N, P, and K) and quadratic saturation D-optimal design was conducted to evaluate the effects of the fertilization rates of N, P, and K on the yield and root baicalin content of Scutellaria baicalensis Georgi at harvesting time. A ternary quadratic polynomial mathematical model was built, in which, the N, P, and K fertilization rates were independent variables, and the yield and root baicalin content were the target functions. Through the analysis of the model, the optimum fertilization pattern was obtained. The results showed that the fertilization rates of N and P had significant effects on the yield of S. baicalensis. Under low fertilization level, the yield increased with increasing N and P fertilization rates; after exceeding definite fertilization range, no obvious effects were observed. N, P, and K fertilization all had significant effects on the root baicalin content. With increasing fertilization rates of N and K, the root baicalin content rates decreased after an initial increase. Under low fertilization level, the root baicalin content increased with increasing P fertilization rate first, and kept stable then. There existed interactive effects between the fertilization rates of N and P, N and K, and P and K on the yield and root baicalin content of S. baicalensis. Under our experimental conditions, the optimum fertilization model for obtaining over 4000 kg x hm(-2) of S. baicalensis yield and > 14% of root baicalin content was 90.5-104.7 kg x hm(-2) of N, 163.9-199.9 kg x hm(-2) of P2O5, and 84.1-140.8 kg x hm(-2) of K2O, with an N:P2O5:K2O ratio of approximately 1:1.86:1.15.