In vegetated areas, due to the interference of high-vegetation cover and numerous shadows, the extraction of alteration information has not achieved satisfactory results. The Zijinshan copper polymetallic ore field in Fujian is located in the southern part of the Wuyishan metallogenic belt, with a high-vegetation cover and many mountain shadows. Conventional methods such as vegetation, water body, and shadow masking combined with principal component analysis have poor performance in extracting mineralization alteration information. In this study, the mixed-pixel decomposition method was used to obtain the reflectance brightness values and abundance information of various end-members of ground objects, and the end-members of interference information such as vegetation were removed. Then, the spectra of rock and soil end-members were reconstructed to eliminate the influence of the vegetation end-member. Finally, the principal component analysis method was combined to extract the alteration information related to mineralization. Based on Landsat 8 multi-spectral data and Zhuhai-1 hyperspectral data, this research adopted the principal component analysis and mixed-pixel decomposition methods to extract the iron-stained and hydroxyl alteration characteristics of the Zijinshan ore field under high-vegetation cover. The results show that the alteration information has a high spatial correlation with known gold deposits such as the Zijinshan high-sulfidation copper-gold deposit and the Yueyang low-sulfidation silver-gold polymetallic deposit. High-value anomalies are displayed at exposed mining areas, ore piles, and tailings slag piles, which is in good agreement with the actual situation.