For pine wilt disease, cutting and fumigation treatment is a general process for infected trees. But time for withering to death after symptoms shown is too short to deter. So techniques for detecting the disease before it spread out are needed. This study shows how Normalized Difference Vegetation Index and Detection Index which is percentage of differences compared to former and latter VI values are suitable for early detection of pine wilt disease and possibility of indexing tree's health. In the areas where pine wilt disease are spread out seriously, images were taken with normal digital camera and CMOS sensor mounted Agricultural Digital Camera to sensing NDVI, in 9 regions, 8 times. After calibration of images, target certainly dead trees in November images and tracking target trees by inverse time series. After tracking of target trees in normal image, same trees are found in NIR images and observed change of vitality of trees. After that, calculated the NDVI values of those trees, and then drew DIs. When DI values are big, it means vitality of tree is lower and lower with time. So when trees show beyond of specific DI value, we doubt the trees getting die for some reasons and inspect those trees, and then treat infected trees in an early stage.