The volatile chemical constituents in traditional Chinese medicine (TCM), Artemisia capillaris herba, were determined by gas chromatography-mass spectrometry (GC-MS). The acquired two-dimensional data were resolved by correlative chemometric resolution methods. The noise in the raw data is pretreated by roughness penalty smoothing method. With the data denoised, heteroscedastic noise and signal-to-noise ratio were decreased apparently, which was favorable to the determination of component number. The selective range can be extracted from rankmap obtained by fixed size moving window evolving factor analysis (FSMWEFA) conveniently. The overlapped chromatographic peaks were resolved into pure chromatograms and pure spectra with evolving window orthogonal projection (EWOP). The purity of the resolved pure spectra were improved furthermore with the heteroscedastic noise decreased through roughness penalty smoothing method, although the basic structure of the raw data changes little. Qualitative analysis was performed by similarity search in NIST147 and Willey library. Pure chromatograms are in favor of quantitative analysis, which was obtained by total volume integration. Forty-two of seventy-five separated constituents in essential oil, accounting for 89.03% of the total content, were identified. The result proves the combined approaches to be powerful for the analysis of complex herbal samples.