Regarding linear estimation theory, the equivalence of the Wiener and Kalman filters is a well-known topic; however, the difference in a practical environment has not been thoroughly discussed. This paper compares the Kalman smoother to the Wiener smoother in terms of practical orthogonal frequency division multiplexing channel estimation on the receiver side. First, conditions for fair comparison are discussed. Under these conditions, the performance and complexity for both methods are numerically investigated. Comparison results show that the Wiener smoother slightly outperforms the Kalman smoother because it avoids cumulative error in sequential processing, while the complexity of the Kalman smoother is always lower than that for the Wiener smoother because there is no large matrix operation.