The major challenging task in multiple input multiple output (MIMO) OFDM systems is to design a good channel estimation method with less computational complexity and lower value of bit error rate (BER). Here, a novel signal processing method based on combining the advantages of multi-scale property of wavelet transform (WT) with blind estimation capability of independent component analysis (ICA), called, a multi-scale ICA (MS-ICA) has been proposed to estimate the channel characteristics. The MS-ICA method can effectively identify the channel impulse response under unknown channel conditions i.e. blindly. Performance of the proposed MSICA method is evaluated by calculating the BER for different fading channels. As compared to other well known OFDM channel estimation methods the proposed MS-ICA method resulted in a lower value of BER, which indicates the superiority of the proposed method. The proposed ASP channel estimation method firstly utilizes the PN based correlation in the time-domain to obtain the auxiliary channel information.