OBJECTIVE To provide a reference for the quality control of Hypericum perforatum. METHODS High- performance liquid chromatography （HPLC） was used to establish fingerprints for 20 batches of H. perforatum and determine the contents of its main components： chlorogenic acid， rutin， hyperin， isoquercitrin， avicularin， quercitrin and quercetin. Cluster analysis was conducted using SPSS 26.0 software. The chromaticity values （luminance value L*， red-green value a*， and yellow- blue value b*） of H. perforatum powder were measured using electronic eye. A prediction model for the contents of seven components in H. perforatum based on its appearance chromaticity values was established using machine learning algorithms. The predictive performance of the models was evaluated using root-mean-square-error （RMSE）. RESULTS A total of 16 common peaks were calibrated in the fingerprints of 20 batches of H. perforatum， and 9 peaks were identified， which were chlorogenic acid， rutin， hyperin， isoquercitrin， avicularin， quercitrin， quercetin， hypericin and hyperforin； the similarities of the 20 batches of samples and reference fingerprint ranged from 0.889-0.987. The contents of chlorogenic acid， rutin， hyperin， isoquercitrin， avicularin， quercitrin and quercetin were 0.025%-0.166%， 0.048%-0.339%， 0.082%-0.419%， 0.017%-0.209%， 0.011%-0.134%， 0.020%-0.135%， 0.041%-0.235%， respectively. Cluster analysis results showed that 18 batches of qualified H. perforatum were grouped into three categories， when the Euclidean distance was set to 1.4. L* of the 20 batches of H. perforatum ranged from 62.814 to 75.668， a* ranged from 1.409 to 3.490， and b* ranged from 25.249 to 30.759. RMSE of three prediction models， namely XGBoost， LightGBM， and AdaBoost， ranged from 0.008 to 0.070， indicating good fitting performance. XGBoost model predicted the contents of the other six components with high accuracy， except for rutin. CONCLUSIONS The established fingerprints and content determination methods are accurate， reproducible， and reliable. The content prediction model based on appearance chromaticity values， combined with machine learning algorithms， can be used for the quality control of H. perforatum.