Abstract : This thesis mainly focuses on the challenging channel estimation problem in the new Chinese digital television (DTV) standard-DTMB. First, we compare DTMB with DVB-T in terms of system specifications, power efficiency and BER. Secondly, we study channel estimation methods exploring the pseudo noise (PN) sequence in frequency and time domains. The frequency domain method is easy to implement. The time domain technique based on correlation of the PN provides a better performance, but suffers from an estimation error floor. Consequently we propose several methods to mitigate the error floor. Finally, we propose new data-aided channel estimation methods which exclude decoding and interleaving from the feedback loop to reduce the computational complexity. Moving average and Wiener filtering in frequency and time domains have been studied to refine the estimates. The new algorithms provide superior performance over the typical methods in the literatures in different applications and channels.