As conventional methods deal with channel estimation and synchronization separately, they can?t achieve the best performances because the unknown parameters will interfere with each other. Besides, the convergence speed is slow using expectation maximization algorithm. Thus in this paper, we propose a joint synchronization and channel estimation method in MIMO-OFDM systems. The algorithm achieves coarse estimation by using a preamble. Then by combining this algorithm with SAGE algorithm and using joint iteration technology, time-frequency synchronization and channel estimation is achieved. Consequently, the performance of the system is enhanced. Theoretical researches and simulation results show that the algorithm can obtain relatively perfect result of time-frequency synchronization and channel estimation with fewer iterative times. Meanwhile our algorithm can approach the ideal case of perfect channel estimation and synchronization with lower complexity. DOI: http://dx.doi.org/10.11591/telkomnika.v11i12.3679