Abstract: This research, by use of RS image-simulating method, simulated apparent reflectance images at senor level and ground-reflectance images of SPOT-HRV, CBERS-CCD, Landsat-TM and NoAA14-AVHRR's corresponding bands.These images were used to analyze sensor's differences cased by spectral sensitivity and atmospheric impacts. The differences were analyzed on normalized difference vegetation index(NDVI). The results showed, the differences of sensors' spectral characteristics cause changes of their NDVI and reflectance. When multiple sensors' data are applied to digital analysis, the error should be taken into account. Atmospheric effect makes NDVI smaller, and atmospheric correction has the tendency of increasing NDVI values. The reflectance and their NDVI of different sewn can be ed to analyze the differences among sensors features. The spectral analysis method based on RS simulated images can provide a new way to design the spectral characteristics of new sensors.