NDVI is an important remote sensing index of characterization of vegetation leaf area density and an important parameter in the estimation of vegetation coverage,forest canopy density,biomass and carbon storage and vegetation growth status and crop yield. Correlation study have indicated that the topography fluctuation has a great influence on the spectral data,but its effect on NDVI is worth further studying. In view of the above problems,the C calibration model was used to correct the topographic correction of Landsat data,eliminating the influence of topographic relief on spectral data,and comparing and analyzing the NDVI. The results show that the terrain correction and overall image of NDVI have significant difference,of which smaller slope vegetation areas have no significant difference,the larger slope vegetation region has significant difference;shady slope overall NDVI tends to be bigger,sunny slope slightly reduced,semi shaded slope has no obvious change;farmland at different slopes has no significant difference and overall change is not obvious,for coniferous forest and broad-leaved forest at shady slope and sunny slope,change of NDVI amplitude is greater than that of half shade and semi sunny slopes. Topographic correction of NDVI estimated value has some influence,thus further influences based on NDVI parameters of various remote sensing estimation results,so eliminating the influence of relief on the spectra can improve the accuracy of quantitative estimation by remote sensing.