Abstract: Sixteen amino acids in 30 batches of Angelica sinensis samples from different producing areas were determined by automatic amino acid analyzer, and the difference of amino acid content and quality of Angelica sinensis were evaluated combining with principal component analysis, cluster analysis and discriminant analysis. The crushed and sieved sample (1 g) was taken, and 30 mL of 0.1 mol·L^-1 hydrochloric acid solution was added. After ultrasound and cooling to room temperature, the volume of the mixture was made up to 100 mL with water. Ailquot (1 mL) was concentrated to nearly dry under vacuum at 50 ℃, and 10 mL of the sample diluent (pH 2.20) was added. The solution was dissolved by vortex, and filtrated by 0.22 μm water filtration membrane. Aspartic acid (Asp), threonine (Thr), serine (Ser), glutamic acid (Glu), glycine (Gly), alanine (Ala), valine (Val), methionine (Met), isoleucine (Ile), leucine (Leu), tyrosine (Tyr), phenylalanine (Phe), lysine (Lys), histidine (His), arginine (Arg) and proline (Pro) were determined according to the working conditions of the instrument. The results showed that the linear ranges of the standard curves of 16 amino acids were 1-200 μmol·L^-1, with detection limits (3S/N) in the range of 0.02-0.04 mg·L^-1, the spiked recoveries in the range of 90.4%-98.2%, and RSDs (n=6) of the determined values in the range of 1.3%-4.1%. The total amounts of 16 amino acids and essential amino acids were in the ranges of 4.86%-10.70% and 0.37%-2.71%, respectively, and the differential amino acids were Gly and Met. Five principal components were extracted by principal component analysis after dimensionality reduction for the contents of 16 amino acids, and the cumulative variance contribution rate was 80.895%. The comprehensive scores of 30 batches of Angelica sinensis samples were in the range of -2.44-2.42, and the highest comprehensive score was sample 13^#, while the lowest score was sample 29^#. On this basis, cluster analysis and discriminant analysis were carried out respectively. The 30 batches of Angelica sinensis samples were grouped into 3 categories by cluster analysis, and sample 28^# (Yunnan) was the first class, sample 29^# (Shaanxi) was the second class, and other samples (Sichuan, Gansu and Qinghai) were grouped together as the third class. The canonical discriminant function models established by discriminant analysis were used to back adjudication verification for 30 samples, and the correct discriminant rate was 100%.