Abstract: In order to explore the change of vegetation in mining area and its influence by mining ac⁃ tivities，aiming at the problem of vegetation classification and monitoring in iron ore area around Anshan， the image data based on Landsat5 TM/Landsat8 OLI is studied.By using the relevant image data，the vege⁃ tation classification of the four major iron ore areas around Anshan in the past 40 years is monitored and analyzed. By using the random forest algorithm，combined with various vegetation spectral indices and tex⁃ ture indices，a random forest tree suitable for the study area is constructed，and the vegetation type classi⁃ fication in this area is completed. The study is divided into six categories: grassland，non-vegetation， shrub，broad-leaved forest，farmland and coniferous forest. The classification and accuracy evaluation are carried out at a time interval of 5 years.The overall accuracy of the research is above 78.69%，and the aver⁃ age Kappa coefficient is 81.97%. The highest accuracy in 1999 is 89.52%，and the lowest accuracy in 2007 is 78.69%，indicating that the overall classification accuracy of the machine learning algorithm is good.The vegetation types in the mining area are mainly transformed from natural grassland to shrub，conif⁃ erous forest and broad-leaved forest.The mining activities in the mining area have a great impact on the surrounding vegetation.

Key words: mining area vegetation, random forest method, remote sensing monitoring, vegetation change