Abstract: Based on Markov Random Fields (MRF) and Gaussian Mixture (GM) models, a new method to label surface features using hyperspectral imaging is presented. The dimension of the hyperspectral image is reduced by PCA, and the stochastic model is built based on prior of the dimension-reduced images and its difference images. Then the maximum posteriori is designed as the optimal criterion and the final labels are obtained by the simulated annealing algorithm. Experimental results show that this method is accurate, efficient and robust for surface features labeling.