目的 采用星点设计-效应面法优选欣宁泡腾片提取工艺。方法 在单因素试验的基础上，选取对提取工艺影响较大的乙醇体积分数、提取时间和溶媒倍量为考察因素，以五味子醇甲、芍药苷、橙皮苷提取转移率为考察指标，对其进行权重计算，综合评定，通过对考察因素各水平进行多元线性回归及二项式拟合，用效应面法选取最佳工艺参数，同时进行预测分析。结果 多元二项式模拟的模型最佳：Y=-209.21+6.67 A+39.68 B+6.59 C+0.077 AB-0.026 AC-0.59 BC-0.051 A²-8.37 B²-0.13 C²（r=0.979 3，P<0.01）。优选的最佳提取工艺条件为：乙醇加热回流提取，乙醇体积分数为65%，提取2次，提取时间为每次2 h，溶媒倍量12倍；在此提取工艺下，五味子醇甲、芍药苷、橙皮苷的平均转移率分别为85.33%、88.77%、93.18%，综合评分为88.72%，与预测值89.22%基本相符。结论 建立的模型可较好地体现因素与指标的关系，具有良好、可靠的预测性。

Objective To optimize the extraction technology for Xinning Effervescent Tablet (XNET) by central composite design/response surface method. Methods Based on single-factor test, the independent variables were ethanol volume fraction, extracting time, and solvent ratio, while the dependent variables were extracting transfer rates of paeoniflorin, hesperidin, and schisandrin, with their weight calculation for a composite score. Multiple linear regression and multivariate binomial fit were used to estimate mathematic relationship between independent variables and dependent variables. Response surface description was based on optimal mathematical model, and then the optimal technical conditions were chosen. Results The best multi-binomial model simulation was as follows: Y = -209.21 + 6.67 A + 39.68 B + 6.59 C + 0.077 AB-0.026 AC-0.59 BC-0.051 A²-8.37 B²-0.13 C² (r = 0.979 3, P < 0.01). Optimal extraction conditions were as follows: ethanol refluxing extraction, ethanol volume fraction of 65%, extraction time of 2 h, and solvent ratio of 12 (mL:g). Under these optimal conditions, the average extracting transfer rates of schisandrinthe, paeoniflorin, and hesperidin were 85.33%, 88.77%, and 93.18%, and the actual extracting rate of catalpol was 88.72%, matching with the predicted extraction rate of 89.22%. Conclusion The established model can better reflect the relationship between the factors and indicators, with good and reliable predictability.