In this paper, we consider the problem of using Space-Time Adaptive Processing (STAP) in ground-based rotating radars, in the case where no jamming alone reference is available. Thus, we propose the use of a STAP algorithm, in which adaptive processing is applied in Beam post-Doppler space. First, we give design rules for the implementation of this processing, after considering two cases depending on the knowledge of the jammers' DOAs and investigating the influence of the number of Doppler filters on performance. Then, we show by simulations that the proposed processing is robust to antenna rotation and has the advantage of having both an high convergence speed with the number of training samples and a low implementation complexity