Imaging Spectrometer is a new developed remote sensing instrument. A newprocessing method of its data-Constrained Spectral Mixture Analysis （CSMA） is presented in this paper. CSMA looks every pixel as mixture of some endmembers. Thecontent of each endmember in a pixel can be calculated using Imaging Spectrometerdata. This is a problem of getting optimal solution of the spectral mixture equation.CSMA gets the optimal solution of the equation with some conditions and a fastalgorithm is proposed here. This result is more interesting than traditional classificationbecause it is more quantitative.