Abstract: Conventional algorithms based on short-time Fourier transform time-frequency analysis can not extract the detailed micro Doppler features caused by motion of different parts of human leg and arm. The fractional Fourier transform (FRFT) is proposed and applied to extract the detail features of gait. The fractional Fourier transform spectrum generated from real measured data is presented and analyzed. Test results show that, by fractional Fourier transform, the detailed micro Doppler features of human limbs swing can be extracted from the experimental gait data.