Abstract: In this paper, an algorithm is proposed to estimate the Linear Frequency-Modulated (LFM) signal's parameters under non-gauss noise condition. By analyzing the characteristics of impulse noise, a modified lower-order-moments ambiguity function is presented, which can analyze the time-frequency trait even under the impulse noise condition. Moreover, the proposed method combines the Radon transform to estimate the LFM signal's parameters, which can suppress noise further. This approach can be applied in the impulse noise case or in the gauss noise case, hence it possesses the better robust property than the conventional methods. At last, the performance of the proposed scheme is compared with the conventional method, and the computer simulation results show the effectiveness of the new method.