Abstract: Super-resolution radar imaging methods break out of Rayleigh Limit,and consequently a remarkable improvement in resolution of radar range profiles favors target recognition.Because conventional schemes lack frequency selectivity,a new super-resolution spectral estimating method based on time-domain pre-filtered ESPRIT technique is studied and successfully applied to radar range imaging.This algorithm combines a frequency-selective time-domain pre-filtered filter with the extension of ESPRIT.Firstly,the prior knowledge of frequency distribution is obtained through a rough estimation of radar echoes,and then an TPF filter designed pertinently is used to pre-filter radar echoes to enhance ESPRIT’s estimating precision and imaging quality at frequency band of radar target around.Imaging results of simulation experiments and darkroom measuring data show excellent frequency selectivity and validity of TPF-ESPRIT.