To reduce the calculation of distributed linearly constrained minimum variance (D-LCMV) beamforming in tree topology wireless sensor network (WSN), a multistage Wiener filter (MSWF) based D-LCMV beamforming method is proposed by introducing a multistage decomposition techniques. By effectively applying the MSWF technology to the new method, the estimation and the inversion of local covariance matrix are avoided. The new method can reach the same output performance as that of the D-LCMV beamformer with less amount of calculation. Therefore it is shown that the new method inherits advantages of both MSWF and D-LCMV beamformer. Simulation results prove that the new method achieves excellent performance.