With the objective of estimating the remaining forest cover on Leyte Island, land cover classifications using Landsat 4 TM, Landsat 5 TM and Landsat 7 ETM+ were performed for the years 1989, 1994, 1996 and 2001. In order to reduce cloud cover and shadow obstruction, a total of 37 images were processed and then fused together. A Support Vector Machine (SVM) was chosen as classifier. The accuracy for the classified images was above 80% for each image. A total of about 66,000 hectares were detected as deforested in the time frame from 1989 to 2001.