In this paper, we investigate and propose channel estimation techniques for multi-carrier code division multiple access (MC-CDMA) system based on pilot arrangement. A low complexity linear minimum mean square error (LMMSE) channel estimation for pilot frequency is proposed. By partitioning channel autocorrelation matrix in LMMSE estimator into small sub-matrices whose size is determined by channel coherence bandwidth, the complexity is signicantly reduced in spite of slight performance degradation. We also propose an improved discrete Fourier transform (DFT)-based interpolation for data subcarriers. It can achieve good estimation accuracy by ignoring non-signicant channel taps which contain more noise than channel power, in which signicant taps are detected by using threshold related to noise power. The proposed algorithms are veried by computer simulation.