Abstract: The acceleration of weak target leads to great spread of Doppler spectrum of the echo and traditional detection methods do not work. To solve the problem, a novel method with surveillance radar is proposed which can detect and estimate the parameters of the accelerating targets effectively. First the energy of chirp signal in slow time domain of the echo is integrated by the fractional Fourier transform to improve the detection performance in low signal to noise ratio(SNR) condition. Then the target parameters are estimated by discrete scaling transform method. Finally, the effect of different factors such as coherent processing interval, fractional step and SNR on detection performance and parameter estimation precision of the proposed method are analyzed by theoretical derivation and simulated experiments.