In the three dimensional reconstruction of cone-beam CT, a fast reconstruction method based on the smallest three dimensional convex hull was proposed for actual scan data. First, according to the definition of the minimum three dimensional convex hull, the minimum three dimensional convex hull of the detected object was obtained in the actual scan using a segmentation algorithm based on the projected images, and then with the Z-line data first algorithm, the image reconstruction area was limited to the minimum three dimensional convex hull to enhance the cone-beam CT reconstruction speed by reducing the redundant computing. The experimental results show that this method can effectively reduce the memory consumption, and significantly improve the reconstruction speed and reduce the noise surrounding the object imaging area.