Orthogonal time frequency space modulation (OTFS) which is a novel modulation scheme, aims at offering reliable communications to users under high mobility conditions. Since each symbol in OTFS experiences the same fading over the doubly selective channel, OTFS is able to exploit full time-frequency (TF) diversity when coupled with a proper equalizer. In this paper, we study the low complexity iterative equalizer design for the coded OTFS system. In order to jointly combat the two-dimensional interference, we adopt linear minimum mean squared error based parallel interference cancellation (LMMSE-PIC) as the equalizer for OTFS. The first order Neumann series is used to approximate the matrix inversion involved in LMMSE-PIC for OTFS, which reduces the equalizer complexity to quasi-linear to the total number of transmitted symbols. Extrinsic information transfer (EXIT) charts and coded bit error rate (BER) simulation results show that our proposed LMMSE-PIC equalizer outperforms the recently proposed PIC approach for OTFS and also achieves considerable gain compared to orthogonal frequency division multiplexing (OFDM).