Abstract: In broadband wireless communication systems, the practicability of MIMO techniques relies on the solutions of multiuser MIMO transmission schemes and feedback schemes of channel state information （CSI）. Most researches on limited feedback focus on point-to-point transmission systems. In multiuser MIMO systems, the reduction of feedback signaling overhead is more important since the base station relies on the CSI of all receiving terminals. In this paper, a transmitter interference elimination （TIE） scheme based on the reverse matrix of signature matrix is proposed to improve the system performance. A user selection algorithm is proposed to choose the transmitting users that generate the minimum interference to each other. A power allocation algorithm based on the water-filling algorithm is also proposed to maximize the system throughput. In addition, a limited feedback scheme based on vector quantization is proposed to reduce the feedback signaling overhead. Finally, the performance characteristics of TIE under limited feedback is studied and proved. It is shown that the multiuser interference generated by quantization error is linearly proportional to the transmitting power. It is also shown both in analysis and in simulation that the performance of multiuser MIMO systems under limited feedback are upper bounded because of the multiuser interference generated by feedback error. Numerical results show that the proposed scheme achieves significant throughput improvement while reducing the feedback overhead.