Forty batches of Lonicerae Japonica Fse i collected extensively and prepared as the test solution. Their chromatographic fingerprints and anti-influenza virus IC50 value (half maximal inhibitory concentration) were determined respectively. Then Unscrambler software was used, and spectrum-efficient correlation analysis was done for chromatographic fingerprints data and IC50 data by partial least squares regression method, to establish spectrum-efficient correlation model for anti-influenza virus of Lonicerae Japonicae Flos. Then the other 10 batches of Lonicerae Japonicae Flos were used to verify the model and explore the adaptability of this spectrum-efficient correlation model based on partial least squares regression method. The mathematical model obtained R2 of 0.969489 and RM-SEC of 0.070691 for calibration set; R2 of 0.959042 and RMSECV of 0.084005 for cross validation set. The verification experiment results showed that the relative error between the predicted values and measured values was within 10% in all 10 hatches, and within 5% in 80% of them. The results showed that the established spectrum-efficient correlation model could be used to evaluate the biological activity of anti-influenza virus of Lonicerae Japonicae Flos by determining its HPLC fingerprints.