Medium resolution images from multispectral sensors like Landsat TM have been extensively used for decades in order to identify decline and defoliation generated by insects and other forest pests. The present work analyses the usefulness of these kinds of images to detect small-size stands of Pinus nigra Arn. and Pinus sylvestris L. affected by Scolytidae attacks. The study area was located in the Solsones region (Eastern Pyrenees), selecting 34 training zones (17 damaged small-size stands and 17 healtly small-size stands). The exploratory analysis of the images was conducted with the ERDAS® IMAGINE 8.x. program.The results of the study showed significant differences between the affected and non-affected stands in 5 of the 7 spectral bands analysed. TM5 and TM7 bands were identified as those having the highest power to detect damaged stands. The digital levels obtained and the spaces of characteristics created, both showed trends to group small-size affected stands versus healthy, achieving improvements in the methodological procedure employed.