A noninvasive and rapid method for the diagnosis of disease was developed on the basis of an analysis of visible and near-infrared (Vis-NIR) spectra from tongue tip. Reflectance spectrum in the 463.87-1737.26 nm region from the tongue tips of 149 volunteers were collected and the samples were separated into two parts: calibration sample and test sample. Spectra were then subjected to two different analysis methods: a partial least squares (PLS) regression analysis and an interval partial least squares (iPLS) regression analysis. PLS and iPLS model gave the best results for test samples with correlation equal to 0.902 and 0.932, and with classification accuracy equal to 75% and 85%, respectively. The results showed that the iPLS method seem more robust than PLS model in full-spectrum region. The result also showed that application of the spectra for disease diagnosis is promising, and may provided a fast and simple diagnostic tools for clinical.