In order to study the optimal extraction technology of essential oil from Amomum tsao-ko by steam distillation and ultrasonic organic solvent extraction.In this paper,based on single factor experiment,the extraction process of A. tsao-ko essential oil by steam distillation and ultrasonic organic solvent extraction was optimized by response surface methodology (RSM) and back-propagation artificial neural network (BP-ANN).Gas chromatography-mass spectrometry (GC-MS) was applied to analyze the essential oil components obtained from the two extraction methods,and the models were compared.The results showed that the optimum technological parameters of A. tsao-ko essential oil extraction by steam distillation were soaking time of 53.00 min,distillation time of 72.00 min,liquid to solid ratio of 11.00 mL/g,and extraction yield of 1.59%.A total of thirty-four components were identified in A. tsao-ko essential oil obtained by steam distillation.Under the optimum conditions of ultrasonic organic solvent extraction with ultrasonic power of 150.00 W,ultrasonic time of 27.00 min and liquid to solid ratio of 6.30 mL/g,the extraction yield was 3.40% and a total of 43 components were identified.Although the extraction yield of A. tsao-ko essential oil by ultrasonic organic solvent extraction was higher than that by steam distillation,the essential oil obtained by ultrasonic organic solvent extraction was poor in quality,dark in color and heavy in impurities,which was not conducive to the subsequent application.Compared with the RSM model,the BP-ANN model had a lower error-index and a higher fitting degree in predicting extraction yield,and had higher accuracy and prediction advantages.This study will provide reliable technological parameters for the efficient extraction of A. tsao-ko essential oil and further promote the development and utilization of the medicinal and edible plant A. tsao-ko.