The feasibility of the determination of the time‐average of pulsatile velocity obtained via a nontriggered magnetic resonance (MR) acquisition is studied. The advantage of this method, in comparison with a triggered acquisition, is a considerable reduction (≈15×) in acquisition time. However, pul‐satility causes image artifacts, known as ghosts, and the Fourier transform technique required for the imaging procedure accomplishes time‐averaging of the complex MR signal. Both effects can result in errors in the velocity determined. Calculations show that these errors depend on the velocity time function and the acquisition parameters. In vivo comparison of triggered and nontriggered MR velocity measurements in the femoral artery of volunteers (n = 7) shows larger statistical and systematic errors in the latter, which depend on the excitation angle. Therefore, this nontriggered average velocity measurement is only useful as a fast and rough estimation of the time‐averaged velocity.