Abstract: Based on the mathematical model of flat fading MIMO channel, the minimum mean square error (MMSE) estimation algorithm was studied with its derivation process given in detail. The LMSE estimation algorithm is so complex that its convergence rate is restricted to a certain extent and becomes slow in applications. To reduce its complexity, the algorithm is simplified mathematically with the inverse operation of first-order matrix cut out. An MIMO system is set up in simulation to provide the curves of estimation performance and those of convergence. The theoretical analysis and simulation results show that no obvious difference is found between the estimated performance of the improved algorithm and that of unimproved algorithm, but the convergence rate of improved algorithm is significantly increased. Thus the simplification is regarded as successful.