Objective Toestablish a scientific and reasonable quality evaluation and grading standard for Armeniacae Semen Amarum (ASA). Methods Taking 32 batches of dried mature seeds of Prunus armeniaca as the research objects, Their morphological parameters (longitudinal diameter, transverse diameter, thickness diameter, kernel shape index), kernel dry weight, hardness, color values (L^*、a^*、b^*、E^*_ab), odor response values, and amygdalin content were determined. Grade classification was performed using the quality constant method,and the correlations between various indicators and quality grades were investigated through correlation analysis, partial least squares-discriminant analysis (PLS-DA), and regression analysis. Results The quality constant A of the 32 batches of ASA kernels ranged from 1.04 to 3.15, Based on the A value, they could be classified into selected grade (six batches) and mixed grade (26 batches). Amygdalin content was significantly positively correlated with hardness and b^* value (P < 0.05), and the quality constant A was extremely significantly positively correlated with dry kernel weight, transverse diameter, hardness, and amygdalin content (P < 0.01). The PLS-DA model could effectively distinguish ASA of different grades, and amygdalin content and quality constant A were identified as the key variables for grading. Regression analysis showed that hardness could explain 20.0% of the variation in amygdalin content, while hardness and transverse diameter together explained 60.10% of the variation in the quality constant. Conclusion The evaluation system constructed in this study based on appearance characteristics, intrinsic components, and quality constant, which helps standardize the market order of ASA, ensure the safety and efficacy of clinical medication. It is of great significance for promoting the standardized development of ASA industry and can provide a scientific basis for the quality evaluation and market grading of ASA.

Armeniacae Semen Amarum  /  quality evaluation through morphological identification  /  appearance traits  /  amygdalin  /  quality constant  /  quality evaluation  /  grade classification