Channel estimation has always been one of the key points in the research of multiple input multiple output (MIMO) communication systems. This paper proposes a hybrid MIMO channel estimation algorithm based on multiple training sequences which combine sequential insertion mode and superposition mode. In this algorithm, a channel estimation model based on multiple training sequences is proposed. Using the periodic superposition of multiple training sequences, the training sequence signal is enhanced and the state estimation of the channel is realized. At the same time, the channel estimation is improved by using the excellent autocorrelation function of the multiple training sequences and the lower peak average power bit. By combining the channel estimation model of the timing insertion mode and the channel estimation model of the superposition mode, the power of the training sequence and the length of the training sequence can be adjusted flexibly, and the spectrum utilization of the MIMO communication system can be improved. The experimental results show that this algorithm is more suitable for flexible channel environment.