In this letter, a reduced-rank beamforming method is presented based on knowledge-aided joint iterative optimization. The proposed adaptive reduced-rank processing is realized by joint optimization for the reduced-dimension matrix and beamforming weight vector. Meanwhile, recursively updating the prior covariance matrix using the spatial spectrum reconstruction technology and the weighted processing improve the estimation precision of the array covariance matrix. The simulation results show that the proposed method is robust to the dimension of the reduced-dimension matrix and significantly improves the output signal-to-interference-plus-noise ratio (SINR) of the adaptive beamformer under the condition of few samples.