<p>The unsupervised fuzzy c-means clustering algorithm has been extensively and successfully applied in hyperspectral image (HSI) analysis for several decades. However, HSI contains a broad spectral range and complex spatial structures, resulting in a diverse pixel distribution. Consequently, traditional variations of fuzzy c-means clustering suffer from low segmentation accuracy. To address this issue, we propose super-aware fuzzy c-means clustering (SAFCM) for HSI segmentation, which innovatively integrates superpixel generation and segmentation into a single objective function. The SAFCM utilizes a spatial distance that combines color intensity and corresponding position coordinates to generate superpixels, which enhances the regularity of superpixels and improves immunity to cluttered pixels. Subsequently, the SAFCM employs superpixel features and corresponding histogram information for target segmentation, which accelerates the clustering process and refines segmentation accuracy. To validate the efficiency and effectiveness of the SAFCM, we conduct experiments on HSI data sets. These results demonstrate that the SAFCM outperforms current state-of-the-art methods.</p>