Correlation analysis of canopy density with remote sensing data for different forest stand is basis for estimating canopy density using remote sensing, which is an important field of forest remote sensing. The relationships of the canopy density with Landsat Thematic Mapper(TM, which includes seven bands represented as TM₁、TM₂、TM₃、TM₄、TM₅、TM₆、and TM₇) data for different forest stand were explored in Shimian County of Sichuan Province of P. R. of China, and how they were influenced by topographically correcting TM using the Lambert Cosine Correction(LCC) model and the Sun Canopy Sensor(SCS) model was also studied here. Firstly, the topographic database and the forest resource GIS database whose data were acquired in1994 in field were created. Secondly, Landsat TM data acquired on June 26,1994 were geometrically corrected by using topographic maps, and matched well with the forest resource database. Thirdly, TM-LCC and TM-SCS were respectively obtained by topographically correcting TM using LCC model and SCS model. Fourthly, the mean and standard deviation of each band of TM, TM-LCC and TM- SCS for each forest sub-compartment were calculated by overlaying the forest resource GIS data with each band of TM, TM-LCC and TM-SCS, and were added into the attribute table of the forest resource database. 1194 sub-compartment samples of relatively lower standard deviation were selected from the forest resource database. Finally, the samples were stratified into eight forest stands according to their dominant tree, and the correlation coefficients of canopy density with each band of TM,TM-LCC and TM- SCS were calculated for each forest stand. It was shown that the correlation coefficients differ along with different band and different forest stand. Correlation coefficients of canopy density with TM2,TM3,TM4,TM5 and TM7 for Tsuga chinensis, with TM4,TM5 and TM7 for Abies fabri, and with TM1 for Picea asperata stand were significant at the 99% level of confidence. The highest is the correlation coefficient of canopy density with TM5 for Tsuga chinensis stand, which is -0.324. The correlation coefficient of canopy density with TM1 for Tsuga chinensis stand is significant at the 95% level of confidence. The correlation coefficients of canopy density with each band of Landsat TM for Betula,Quercus,Alnus cremastogyna, soft broadleave and Pinus yunnanensis stand were not significant at the 95% level of confidence. The correlation coefficients of canopy density with TM₄ and TM₅ for Tsuga chinensi, Abies fabri and soft broadleave were enhanced by topographically correcting TM4 and TM5 using the LCC model, which are respectively -0.394,-0.374,-0.209,-0.210,0.545 and 0.577, and significant at the 99% level of confidence. The correlation coefficient of canopy density with TM7 for soft broadleave was enhanced by topographically correcting TM7 using the LCC model(from 0.051 to 0.513), and significant at the 99% level of confidence. The correlation coefficients of canopy density with TM4 and TM5 for Abies fabri was enhanced by topographically correcting TM4 and TM5 using CSC model (from -0.170 to -0.213 and from -0.181 to -0.207), and significant at the 99% level of confidence . The correlation coefficients of canopy density with Landsat TM for Betula,Quercus,Alnus cremastogyna, Pinus yunnanensis and Picea asperata were not significantly enhanced by topographically correcting Landsat TM using the models of LCC and SCS. The study is of important value to stand canopy density remote sensing.