A subtalar joint information extraction method based on images is presented.Talus and calcaneus were segmented from the images by using maximum between-class variance thresholding method combined with interactive morphological method.Subtalar joint complex was reconstructed by marching cubes surface reconstruction method and structures of talus and calcaneus were shown by their animations.A method of subtalar joint information extraction,i.e.principal axes method,was presented to figure out the rotational directions of talus and calcaneus and the angles between them.The results of comparison with another algorithm and verification by simulative model show that the method has high veracity and precision.