Traditional Chinese medicine (TCM) preparations have become effective treatments for many diseases. However, their active ingredients are still uncertain and difficult to identify. In this study, we propose a strategy that integrates ultra-performance liquid chromatography/quadrupole-time-of-flight mass spectrometry (UPLC/Q-TOF-MS) and bioactive (NF-κB inhibitor) luciferase reporter assay systems for the rapid determination of various anti-inflammatory compounds of TCM preparations. In this way, Bufei Granule (BFG), a TCM preparation used for the clinical therapy of asthma, was analyzed by the two ways of component identification and activity detection. Potential anti-inflammatory constituents were screened by NF-κB activity assay systems and simultaneously identified according to the mass spectrometry data. Three structural types of NF-κB inhibitors (caffeic acid derivatives, flavonoids and Pentacyclic triterpenes) were characterized. Further cytokine detection confirmed the anti-inflammatory effects of the potential NF-κB inhibitors. Compared with conventional chromatographic separation and inhibitory activity detection, integrating UPLC/Q-TOF-MS identification and virtual validation was more convenient and more reliable. This strategy clearly demonstrates that MS data-based fingerprinting is a meaningful tool not only in identifying constituents in complex matrix but also in directly screening for powerful trace ingredients in TCM preparations. Copyright © 2016 John Wiley & Sons, Ltd.