PURPOSE: To determine the best MR sequence for evaluation of the anatomical structures of normal kidney. MATERIALS AND METHODS: Twenty normal volunteers (M:F=15:5) took part in this study, and for each, seven sequences were performed. The T1 weighted sequences were conventional spin echo T1 (Conv-SET1), turbo spin echo T1 (TSET1), and fast low angle shot (FLASH), while the T2 weighted sequences were turbo spin echo T2 (TSET2), half-Fourier acquisition single-shot turbo spin echo (HASTE), true-fast imaging with steady-state precession (True-FISP), and echoplanar imaging (EPI). The study involved quantitative and qualitative analysis. In quantitative analysis, CNRs between cortex and adjacent fat tissue, and between cortex and medulla were calculated from SNR (signal to noise ratio), and the CNRs of sequences were statistically compared. In quantative analysis, three radiologists collectively evaluated kidney outline, corticomedullary division, the renal vessels, the pelvis/ureter, and artifacts. For each sequence a grade was assigned, and for each parameter the grades were compared. RESULTS: Between cortex and adjacent fat, the highest CNR was shown by TSET1, followed by Conv-SET1,while among T2 sequences, the CNR shown by TSET2 was highest. Between cortex and medulla, the CNR demonstrated by the three T1 sequences showed no statistically significant difference. Among T2 sequences, however, HASTE showed the highest CNR, followed by EPI, and statistically, the findings for these two were significantly different from those of other T2 sequences. Among T1 sequences, FLASH provided the best kidney outline, though among T2-sequences there was no statistically significant difference. FLASH was also the best for cortico-medullary distinction, while for this purpose the best T2 sequence was HASTE. True-FISP was best for the evaluation of renal vessels, and HASTE for evaluating the pelvis and ureter. Artifacts were most prominent on Conv SET1. CONCLUSION: For evaluating the shape of the kidney, the best T2 sequence was TSET2, but the best T1 sequence could not be determined. For cortico-medullary differentiation, the best T1 sequence was FLASH and the best T2 sequence was HASTE. For the evaluation of renal vessels, True-FISP was best, and for the pelvis and ureter, HASTE. Artifacts were most prominent on Conv-SET1.