The traditional forest protection and statistics in karst area mostly rely on a large number of manual mapping and high resolution image visual interpretation, which not only consumes a lot of time and effort, but also can not guarantee the accuracy of the final results. In order to make up for this subjective error, introduces multi-source data fusion of 10 m resolution hyperspectral satellite image and 0.8 m resolution high resolution satellite image,and combines with the spectral acquisition method of ground object samples, to study the forest classification of Wumengshan Geopark in Liupanshui area of Guizhou Province,The results show that the classification accuracy of SAM based on hyperspectral and high-resolution fusion image data is higher than that of single data source, and the classification accuracy based on fusion image is 85.9%, which basically meets the requirements of surface feature classification and application The precision requirement of adjustment and drawing.