This thesis presents the author's research work on the FDK-type CT imaging algorithms for 3D cone-beam CT scanner. The modern FDK-type algorithms suffer from two main limitations of the data collection: the cone-shaped projection volume and the helical scanning pitch. Once the cone-beam angle becomes relatively large or the helical scanning is relatively fast, the artifacts would unavoidably appear on the images reconstructed by the existing FDK-type algorithms. On the other hand, the FDK-type approximate reconstruction accuracy is to be improved if compared with the 3D exact reconstruction. The research work presented in this thesis is on the development of new FDK-type reconstruction algorithms. Author’s efforts are mainly on the improvements of approximate reconstruction geometry and the incorporation with the 3D exact reconstruction. New methods can produce images with improved quality and better reconstruction efficiency.