This paper builds upon the original 3D statistical model-based iterative reconstruction algorithm developed for CT systems with a flying focal spot. The proposed method adopts a continuous-to-continuous data model and formulates the reconstruction task as a shift-invariant system. While conceptually similar to the FDK algorithm, our approach offers significant improvements in image quality compared to traditional filtered backprojection (FBP), enabling a potential reduction in the patient’s x-ray dose. A key advantage of the method is its direct use of projection data without requiring interpolation, resulting in high-resolution images and improved computational efficiency. Unlike nutating schemes commonly used in dual-source CT systems, this approach avoids their inherent complexities. Simulation results demonstrate that the proposed method not only surpasses standard FDK in image quality but is also competitive in terms of computational time.