This paper investigates training based channel estimation for multiple input multiple output (MIMO) system. In the channel estimation process, least squares (LS), minimum mean square error (MMSE), scaled least square (SLS) and relaxed mean square error (RMMSE) techniques are considered. Performances of all of these techniques are compared while the BPSK signals from multiple antennas are transmitting over a real time channel like IEEE 802.11n WLAN channel. During performance evaluation, the normalized mean square error (NMSE) is computed between actual channel and estimated channel. In simulation analysis, NMSEs of all mentioned estimation techniques are compared and it was found that the RMMSE technique is better among all. © Research India Publications.