目的 采用星点设计-效应面法优化穿心莲的超声提取工艺。方法 以乙醇浓度、提取时间和溶媒比为自变量，以穿心莲内酯、新穿心莲内酯、脱水穿心莲内酯的收率和浸膏得率为因变量，通过对自变量的总评归一值进行多元线性回归和二项式拟合，采用效应面优化最佳的工艺条件。结果 确定超声提取工艺条件为乙醇浓度(X₁：70％～80％；提取时间(X₂：50～70 min；溶媒比(X₃：10～12倍，工艺验证的总评归一预测値与实测值的偏差为-2.6%，二项式拟合的复相关系数r=0.971 3。结论 采用星点设计-效应面法优选方法优化的穿心莲内酯提取工艺，方法简便、预测性良好。

OBJECTIVE To optimize the process of extracting effective constituents from Andrographis paniculata by central composite design-response surface methodology. METHODS Ethanol concentration， time for ultrasound extraction and ratio of solvent to raw material were selected as independent variables. The overall desirability of the extraction rates of andrographolide， neoandrographolide， and dehydroandrographolide and the extract yields were set as dependent variables. Linear or nonlinear mathematic models were used to estimate the relationship between independent and dependent variables. Response surface methodology was used to optimize the process of extraction. RESULTS The multiple correlation coefficient of binomial fitting of the complex model was as high as 0.971 3. The optimal conditions of extraction process were determined as follows： The materia medica was ultrasound extracted for 50-70 min with 10-12 fold solvent of 70%-80% ethanol. The deviation between observed and predicted values was -2.6%. CONCLUSION The central composite design-response surface methodology is convenient and highly predictive for optimizing the extraction process of Andrographis paniculata.