In recent years, spectral imaging technique has been applied widely in mineralogy and petrology. The technique combines the spectral technique with imaging technique. The samples can be analyzed and recognized both in spectra and space by using the technique. However, the problem is how to acquire the needful information from a large number of data of spectral imaging, and how to enhance the needful information. In the present paper, the experimental data were processed by using the technique of data fusion of microscopic spectral imaging. The space distribution map of chemical composition and physical parameters of samples were obtained. The result showed that the distribution of different hydrocarbon in the reservoirs, pore connectivity, etc. were revealed well. The technique of data fusion of microscopic spectral imaging provided a new method for reservoir characterization.