Reclaiming steep slope land to extend the area of slope fields plays a very important role in soil and water loss and eco-environment deterioration in the Loess Plateau.In order to convert farm land into forest on slope land effectively, obtaining up-to-date and reliable information on the spatial distribution and regional extent of slope fields by remote sensing is of critical importance.Using Landsat TM 5 data of the loess hilly and gully area of northern Shaanxi Province, integrated supervised and unsupervised classification were applied to extract slope field and other categories. By improving the signature selection accuracy, this method improves classification accuracy greatly. The result show s that integrated classification is suitable to slope field investigation in the loess hill and gully area.