OBJECTIVE： To study the spectrum-effect relationship of the antipyretic effect of Radix Bupleuri injection. METHODS： HPLC method was used to establish the fingerprint of Radix Bupleuri injection. The antipyretic effect of 10 batches of Radix Bupleuri injection on fever rats induced by dried yeast were determined respectively. The fingerprint peaks were screened， along with the temperature value of different time points （0， 10， 15， 30， 40， 55， 70 min）. The principal components were extracted by principal component method and the indirect relationship between the fingerprint and antipyretic effect was analyzed， depending on the principal component correlation coefficient matrix. RESULTS： There were 39 common peaks in fingerprint （similarity＞0.85）， No. 8， 12， 14， 19， 26， 31， 34， 35 and 39 common peaks with large peak area were included in the study. Four principal components were extracted by principal component analysis （87％ of total variant）. First principal component showed that the active components of the 12th peak may be related to the antipyretic effect of 6 to 13 hours. The second principal component showed that the active components of the 26th peak may be related to the antipyretic effect of 0.5 to 5 hours. The third principal component showed that the similar effect of the active components could be caused by 34th， 35th and 39th peaks. The fourth principal component suggested that there were some similarities between the 14th and the 31st peaks. CONCLUSIONS： Radix Bupleuri injection have obvious improvement for fever rats. There is certain corresponding relation between HPLC fingerprint and antipyretic effect of Radix Bupleuri injection .