Signal-to-noise-ratio (SNR) is an important index for evaluating multispectral imaging performance. In the early stage of designing a multispectral imager, SNR analysis should be performed for the determination of subsystem parameters. The imaging chain model of multispectral remote sensing is introduced. Characteristics of radiation source, object spectral reflectance, atmospheric radiation transfer, optical imaging, dispersive element, detector spectral responsivity and camera noise are synthesized to come into an end-to-end analysis. A filter array based multispectral imaging system is modeled. Computation of atmospheric radiation transfer is from MODTRAN under conditions of different solar zenith angles and different surface features. The models of charge coupled device are adopted for characterizing the noise features. SNR for different working conditions are shown for the designed multispectral remote sensing system. The SNR analysis is helpful for demonstrating the best operation condition for multispectral remote sensing system of this kind, as well as optimizing related optical system parameters for best performance with lowest cost.