Abstract: Abstract: Considering that the existing interpolation parameters estimator for linear frequency modulation (LFM) signals based on fractional Fourier transform (FrFT) declines in performance, and it is not solid for LFM signals with different parameters, a modified interpolation parameters estimator is investigated. Firstly, the causes of performance reduction in the existing algorithm are analyzed, then we define the discrete frequency in fractional Fourier domain. We find that the existing interpolation parameters estimator will make up a reversed compensation when the initial frequency of LFM signal is near to the discrete frequencies under the condition of low signal to noise ration(SNR). The estimator performance will reduce quickly because of reversed compensation. The modified interpolation estimator employs a new method to determine the sign of compensation to improve the immunity to noise, then in an effort to make performance better, we shift the initial frequency of LFM signal to avoid nearing to discrete frequencies. At last, the simulation results indicate that the modified interpolation estimator do well in parameters estimation and is steady for different LFM signals.