In the downlink of a multi-user MIMO system, an appropriate pre-coding algorithm is needed to suppress co-channel interference. Based on the criterion of maximum signal-to-leakage-and-noise-ratio (SLNR), a successive iterative optimization pre-coding scheme is proposed. The method uses the unused subspace of known users to improve the space degree-of-freedom of the other users, and reduces the leakage. An iterative optimal method is applied to further suppress multi-user interference (MUI) and improve the signal-to-interference-and-noise-ratio (SINR) at the receiver. The proposed scheme only makes use of the channel information of the users at the transmitter to optimize the pre-coding matrix without collaboration between the base stations and the users. Any detection algorithm can be selected at the receiver without increasing the system overhead. Simulation results show that the proposed scheme has better performance and faster convergence than the original SLNR algorithm, and therefore is suitable for practical applications.