In modern wireless communication systems, Channel State Information(CSI) is mostly estimated at the receiver and fed back to the transmitter. Unfortunately，the CSI arriving at the transmitter would be outdated due to feedback delay, which may cause significant performance degradation. A trade-off way is proposed，which exploits the time, space, frequency correlation effectively while maintaining a low complexity. First，it is analyzed that the channel correlation can be decoupled into time, space，frequency part. Then, three separate 1-D filters are used which consider time, frequency，spatial correlation respectively based on the separation property of the channel correlation function. Simulation results show that a trade-off between an attractive prediction performance and computational complexity in correlated MIMO-OFDM systems is realized in the proposed algorithm.