Principal Components Analysis "PCA" is a statistical technique that transform a multi-variate data set consisting of inter-correlated variables into a data set consisting of variables that are uncorrelated linear combinations. The transformed variables are referred to as Principal Components PC's. PC *s are chosen in such a way that the first PC expresses the maximum possible proportion o the variance in the original data set. One of the applications o/PC in remote sensing problem is detecting and monitoring tempor changes. In our project the PCA of six channel data sets consisting of multi-temporal Landsat TM images at different times 1988 and 1990 have been studied. The study area has been flooded in 1988, while in 1990 it becomes completely dry In this thesis the PCA has been applied to the six bands in two dates Landsat TM data space inorder to determine the changes that occur due to the flood. We have applied PCA to the images that obtained from the subtraction of the 1990 and 1988 images, also we have applied the PCA to the images obtained from the ratio of bands i.e. the images resulted from dividing the images 1990 by images of 1988. A hybrid method was adopted in our work. In this method the PCA applied to bands 1, 5 and 7 in 1988 and bands 1, 5 and 7 1990, separately for each year then we get the differences between the images by subtracting the corresponding PC's images Our results shows that the best method for studying the changes is the Principal Component applied to the image ratio (PCR) at different time. Different color models was applied to the original bands and to the result obtained for the PC's image. These color models are (YIQ, XYZ, and HIS). These models give good result to monitor the flooded area.