Objective:To detect the features of the teeth image effectively,SIFT algorithm was preformed to use in the image of teeth model. Methods:This approach identified potential interest points by searching over all scales and image locations with the difference-of-Gaussian function. Keypoints were selected based on their stability and a detailed model was obtained to determine location and scale of the candidate location. Assign one or more orientations to each keypoint location based on local image gradient directions. The gradients of local image in region around each keypoint were measured,and the features were matched by KNN algorithm. Results:This method can detect the features effectively and offer some available parameters for 3D reconstruction of the teeth model through lots of experiments and comparing with other feature extraction algorithms. Conclusion:SIFT algorithm performed in the image of teeth model may contribute to 3D reconstruction of the teeth.