Heqing county is located in northwestern Yunnan Province, the southwestern margin of the Yangzi platform, E 100°01′～100°29′ and N25°57′～26°42′, the southern end of the traverse mountains in western Yunnan and the east of the Yunling range. Karst mountainous plateaus and basin landforms dominate this region with a total land area of 2,323.54 km2, including exposed carbonate rock 741 km2, representing typical karst rocky desertification (KRD for short). The research on KRD is of great significance to soil erosion management and improvement of geological and ecological environments in Heqing county. The KRD interpretation based on high-resolution remote sensing images is one of the important methods in large-scale investigations to this phenomenon. At present, visual interpretation is still the main approach for KRD interpretation with high-resolution remote sensing images. To improve the accuracy and speed of the KRD interpretation, a method of band ratio and density slice for KRD investigation is proposed. This study took Heqing county as an example and RapidEye multispectral remote sensing image of 2011 as information source. On the basis of standard spectral curves of limestone and conifer-meadow mix, and combining with analyzing profile curves of different rock coverage DN values, we applied the band ratio and density slice method to investigation of large-scale KRD in Heqing county. After indoor interpretation, we went to the field to validate the accuracy of the interpretation results and modify them. At last, we obtained the KRD data of Heqing county. The investigation results show that the area of KRD in Heqing county is 188.82 km2, which accounts for 8.13% of the whole county area and 25.48% of the karst area. The area of severe KRD is 10.31 km2, which accounts for 5.46% of the total KRD area. The moderate KRD area is 42.28 km2, which accounts for 22.39% of the whole. And the mild KRD area is 136.23 km2, accounting for 71.15%. Through the field validation, the correct ratio of KRD interpretation is 84%, which can meet the interpretation demand and the technical specification. So it is concluded that the band ratio and density slice method can be applied to large-scale KRD remote sensing investigation effectively.