In shallow water areas, to enhance underwater targets detection performance improve computation efficiency of active sonar, a computationally efficient adaptive beamformer (spatial filter) based on inverse QR (IQR) and recursive least-squares (RLS) is developed under fast Fourier transform framework, for standard hexagonal receiving array implementation. The IQR-RLS algorithm has good numerical stability and can be mapped onto coordinate rotation digital computer processor-based systolic arrays, which is suitable for real time applications. Using the proposed scheme to construct beamformer, which reduces computational complexity significantly and offers better converge rate than conventional adaptive beamformer. The simulation and lake test results demonstrates the algorithm improves interference (reverberation) suppression ability. It improves SNR about 2dB of still bottom target detection in reverberation limited area.