Head and neck tumor is a common malignant tumor with increasing incidence and mortality worldwide. Traditional diagnostic methods for head and neck tumor are limited by the lack of specific biomarkers and the limitation of invasive detection methods， as well as high time cost and a high rate of misdiagnosis. Therefore， research on head and neck tumor diagnosis based on new technologies is needed. Hyperspectral imaging （HSI） is a non-contact optical imaging technique that captures a series of images in multiple spectral bands to generate a hyperspectral image cube. HSI has shown its corresponding potential in the early diagnosis， tumor margin identification and clinical research of head and neck cancer.