Abstract: OBJECTIVE To investigate the optimum moulding process of total flavones of Bidens bipinnata dropping pills. METHODS Independent variables were the proportion of substrate to drug, dropping distance, and Tween 80 content, while dependent variables were pills weight variation, roundness, limited disintegration time and overall desirability. The central composite design-response surface method was used to optimize the optimum moulding process. Multivariate linear regression and second-order quadratic models were fitted to establish the relationship between independent and dependent variables. The optimum moulding process was predicted by response surface method. And model prediction and confirmatory tests were carried out in the end. RESULTS The optimum moulding process of the optimization were proportion between drug and substrate 1.8 mL∶1 g, dropping distance 2.20 cm, and Tween-80 content 3.18%. Bias between observed and predicted values of all inspection indexes were small. CONCLUSION The central composite design-response surface method are useful for the optimization of moulding process of flavones of Bidens bipinnata dropping pills. The estimation of the established model is good.