In this paper, a low complexity algorithm is proposed for dual pseudorandom noise (PN) time domain synchronous OFDM (TDS-OFDM) systems, to simultaneously estimate the carrier frequency offset (CFO) and the sampling frequency offset (SFO). The second PN of every OFDM frame is used in the estimation. Firstly, SFO is estimated under maximum likelihood (ML) criterion. After that, a closed-form solution of CFO is obtained. Because the proposed algorithm does not change the original frame structure, there is no need to insert extra symbols when estimating frequency offset. Simulation results show that the proposed algorithm performs slightly better than existing ML estimator, but complexity is much lower. And the performance of the proposed algorithm is different with different PN lengths.Index Terms-time domain synchronous OFDM, carrier frequency offset, sampling frequency offset, dual pseudorandom noise