A modified per-survivor iterative timing recovery (MPS-ITR), which jointly performs timing recovery, equalization, and error-correction decoding, has been proposed in [I] to deal with the problem of timing recovery operating at low signal-to-noise ratio. Practically, this scheme exploits a split preamble strategy in conjunction with a per-survivor soft-output Viterbi algorithm equalizer to make it more robust against severe timing jitters or cycle slips. Although the MPS-ITR outperforms other iterative timing recovery schemes [I], it still has very high complexity. In this paper, we propose a reduced-complexity MPS-ITR scheme (denoted as MPS-ITR-M) to make it more implementable in real-life applications. This is achieved by applying the M-algorithm [2] to the MPS-ITR. Numerical results indicate that at low-to-moderate complexity, the MPS-ITR-M will perform better than other schemes.