摘要Objective: The Box-Behnken response surface method combined with fingerprints was used to optimize the extraction process of total anthraquinone from Cassia seeds. Methods: A three-factor, three-level response surface test was conducted based on the single-factor test with comprehensive evaluation as the measurement index. The comprehensive evaluation indexes included the extraction rate of total anthraquinone of Cassia seeds or the equivalent amount of herbs per gram of total anthraquinone of Cassia seeds, the normalized value of peak areas of 5 index components such as aurantio obtusin in the fingerprint of each sample to 16 shared peaks and the similarity of fingerprints (the reference fingerprint was established by the extraction solvent for the determination of Cassia seeds content in the Chinese Pharmacopoeia 2020 edition) with the weights of 0.2, 0.5 and 0.3, respectively. Results: The best extraction process was obtained: the liquid-to-material ratio was 20:1 (mL g-1), the extraction solvent was mixture of 60％ ethanol-ethyl acetate (2:1), and the extraction time was 15.12 min. The results of five sets of validation experiments showed that the overall evaluation of total anthraquinone of cassia seeds by the best process was 0.528 (RSD=0.45％), and the prediction result of response surface method model was 0.531, and the relative error with the prediction result was 0.531. The relative error of the predicted results was 0.56％, and the best extraction process was consistent with the model prediction, and the obtained best process could be used for the extraction of total anthraquinone from Cassia seeds. Conclusion: The Box-Behnken response surface method combined with the fingerprint technique can be used to find the best reaction conditions and examine the interactions among the factors in a comprehensive and accurate manner, which can provide reference for the optimization and evaluation of the extraction process of Chinese medicine.