objective: To evaluate the image quality of CT urography (CTU) obtained with ultra-high-resolution CT (U-HRCT) reconstructed with hybrid iterative reconstruction (IR) and model-based IR algorithms. methods: Forty-eight patients who underwent CTU using the U-HRCT system were enrolled in this retrospective study. Excretory phase images were reconstructed with three protocols: Protocol A: 1024-matrix, 0.25 mm-thickness, and model-based IR; Protocol B: 1024-matrix, 0.25 mm-thickness, and hybrid IR; Protocol C: 512-matrix, 0.5 mm-thickness, and model-based IR. Objective image noise and contrast-to-noise ratio (CNR) of the renal pelvis were compared among the protocols. Three-dimensional maximum intensity projection CTU images were generated from each image data set, and image quality was evaluated by two radiologists. results: Protocol C yielded the lowest objective image noise and highest CNR, whereas Protocol A had highest image noise and lowest CNR (P <  0.01). Regarding the detailed delineation of urinary tract structures on the images, the mean visual score was significantly higher for Protocol A than for Protocols B and C (P <  0.001), and the mean score for subjective image noise was significantly lower for Protocol A than for Protocols B and C (P <  0.001). conclusions: CTU with a 1024-matrix and model-based IR depicted the structures of the urinary system in the most detail.