Abstract: To meet the demands of automatic pruning machines, this paper presents a new method for dynamic automatic identification of standing tree limbs and capturing the digital images of Platycladus orientalis using the methods of computer vision, image processing, wavelet analysis technology, compressing, filtering, segmentating, denoising and capturing the edge of the picture.Then the authors present an arithmetic for dynamic automatic identification of standing tree limbs, extracting the basic growth characteristics of the standing trees such as the form, size, bend degree and relative space position.That is to use the pattern recognition technology to confirm the relative proportionate relationship matching the database, then achieve the goal for dynamic automatic identification of standing tree limbs.