Time-interleaved analog-to-digital converters (TIADCs) are widely used for multi-Gigabit orthogonal frequency division multiplexing (OFDM) based systems because of their attractive high sampling rate and high resolution. However, when not perfectly calibrated, mismatches such as offset mismatch, gain mismatch and timing mismatch between parallel sub-ADCs can significantly degrade the system performance. In this paper, we focus on offset mismatch. We analyze two calibration techniques for the offset mismatch, based on the least-squares (LS) and linear minimum mean-squared error (LMMSE) algorithms assuming an AWGN channel. The simulation results show that our method is capable of improving the BER performance. As expected, the LMMSE estimator outperforms the LS estimator. However, at large offset mismatch levels or low noise level, both estimators converge. In this paper, we derive the condition on the mismatch level for convergence between the two estimators.