Abstract: A novel algorithm for high-speed multi-target detection and parameters estimation with narrowband radar is proposed in this paper. Firstly, base on cyclostationarity, moving parameters are roughly estimated with the joint Keystone transform and ambiguity searching. And then, according to the coarse estimation, the joint compensating range migration in frequency domain and fractional Fourier transform are utilized to conduct the targets detection and parameters estimation. In the case of multi-target, range migration, Doppler spread and Doppler ambiguity, this algorithm is suitable, and retains the merit of low computational complexity of cyclostationarity. Compared with existing cyclostationarity based algorithms, the weaknesses of low estimation precision and limited range of moving parameters are conquered in engineering applications. The validity of the proposed algorithm is demonstrated by computer simulation and raw radar data results.