The aim of the present study is to develop a submillimeter volumetric (three‐dimensional) fluid‐attenuated inversion recovery sequence at 7T. Implementation of the fluid‐attenuated inversion recovery sequence is difficult as increased T1 weighting from prolonged T1 constants at 7T dominate the desired T2 contrast and yield suboptimal signal‐to‐noise ratio. Magnetization preparation was used to reduce T1 weighting and improve the T2 weighting. Also, practical challenges limit the implementation. Long refocusing trains with low flip angles were used to mitigate the specific absorption rate constraints. This resulted in a three‐dimensional magnetization preparation fluid‐attenuated inversion recovery sequence with 0.8 × 0.8 × 0.8 = 0.5 mm3 resolution in a clinically acceptable scan time. The contrast‐to‐noise ratio between gray matter and white matter (contrast‐to‐noise ratio = signal‐to‐noise ratio gray matter − signal‐to‐noise ratio white matter) increased from 12 ± 9 without magnetization preparation to 28 ± 8 with magnetization preparation (n = 12). The signal‐to‐noise ratio increased for white matter by 13 ± 6% and for gray matter by 48 ± 15%. In conclusion, three‐dimensional fluid‐attenuated inversion recovery with high resolution and full brain coverage is feasible at 7T. Magnetization preparation reduces the T1 weighting, thereby improving the T2 weighted contrast and signal‐to‐noise ratio. Magn Reson Med, 2010. © 2010 Wiley‐Liss, Inc.