Abstract: With the application of active phased array and digital beamforming techniques, low earth orbit(LEO) satellite communication array antenna not only forms fixed multiple transmit and receive beams to cover the visible earth, but also has the ability to dynamically adjust the pattern according to the requirements of communication task. The existing high performance pattern synthesis optimization algorithms can not be directly used in space-borne online real-time computing. In this paper, an efficient iterative least-squares pattern synthesis method is proposed with simpler iteration process and less adjustable parameters compared to the most popular iterative algorithms under the same synthesis accuracy and sidelobe level requirements. Pattern approximation with weighted regions and pattern phase constraint relaxation are adopted. The proposed method can partially meet the online weight vector update requirement for satellite communication pattern dynamic synthesis.