With the continuous development of research work on next generation wireless communication, millimeter wave (mmwave) and Massive MIMO technology has become the focus topics both at home and abroad. In this paper, we choose dense urban environments as our research background, according to the main lineofsight (LOS) propagation and the excellent reflection performance in millimeter wave multiuser Massive MIMO systems, we construct a simplified system model. Based on the sparsity of the beamspace channel matrix, we choose several beams that capture the channel’s main energy to approximate the original highdimensional channel, and propose a lowdimensional beamspace precoder method. The simulation results demonstrate that the reducedcomplexity precoder can also achieve considerable system sum rates. The construction of the system model and the optimization of precoder methods have strong theory value and practical significance, can provide a reference for hardware and software design and subsequent algorithm research.