Channel estimation schemes for fixed wireless access (FWA) multiple-input, multiple-output (MIMO) systems are considered in this study. The use of multiple antennas in combination with advanced detection techniques, such as turbo equalization is an effective means for a FWA system to provide high quality and high data rate services. Accurate knowledge, i.e., a good estimate of the underlying channel is essential for turbo equalization to achieve good performance. In this paper, we investigate some algorithms that are suitable for estimating FWA MIMO channels. The proposed schemes are evaluated and compared using different training sequences. Based on our analysis and numerical results, some recommendations are made on how to design appropriate channel estimator and how to choose training sequence for practical FWA systems.