The SPME-GC-MS method is applied to analyze the characteristic components of combustion soot generated by four kinds of petroleum chemicals fuels i. e.the gasoline

the diesel

the polystyrene and the ABS (acrylonitrile butadiene styrene).The spectrums of these four different combustion soot and the corresponding characteristic components are obtained by this method.Based on the retention time and the corresponding characteristic components summarized from the spectrums of the four kinds of combustion soot

a 20 × 35 mathematic matrix is obtained.The PCA method is applied to analyze this matrix and in succession three main principal components are summarized

with the cumulative rate up to 92%

which indicates that the two or three dimensional coordinate system composed of the principal components is quite effective to classify the combustion soot of different fuels.Finally

based on the results of PCA

an advanced pattern recognition method i. e.the HCA is applied to classify the four types of combustion soot with 100% accuracy.