Orthogonal frequency division multiplexing (OFDM) is well-known for its efficient high speed transmission and robustness to frequency-selective fading channels. On the other hand, multiple-input multiple-output (MIMO) antenna systems have the ability to increase capacity and reliability of a wireless communication system compared to single-input single-output (SISO) systems. Hence, the integration of the two technologies has the potential to meet the ever growing demands of future communication systems. In these systems, channel estimation is very crucial to demodulate the data coherently. For a good channel estimation, spectral efficiency and lower computational complexity are two important points to be considered. In this thesis, we explore different channel estimation techniques in order to improve estimation performance by increasing the bandwidth efficiency and reducing the computational complexity for both SISO-OFDM and MIMO-OFDM systems. We first investigate pilot and Expectation-Maximization (EM)-based channel estimation techniques and compare their performances. Next, we explore different pilot arrangements by reducing the number of pilot symbols in one OFDM frame to improve bandwidth efficiency. We obtain the bit error rate and the channel estimation performance for these pilot arrangements. Then, in order to decrase the computational complexity, we propose an iterative channel estimation technique, which establishes a link between the decision block and channel estimation block using virtual subcarriers. We compare this proposed technique with EM-based channel estimation in terms of performance and complexity. These channel estimation techniques are also applied to STBC-OFDM and V-BLAST structured MIMO-OFDM systems. Finally, we investigate a joint EM-based channel estimation and signal detection technique for V-BLAST OFDM system.