In this paper, a new subarray least mean square algorithm (SLMS) for digital beamforming antenna system is proposed, which is suitable to be applied in parallel DSP hardware to achieve high computational efficiency. Low sidelobe of the beam is obtained by constraining the static steering vector with Tschebyscheff coefficients. Simulation results show that the proposed SLMS method requires less computational load than that of subarray asynchronous recursive least square (SARLS) method.