In cellular networks, when multiple base stations (BSs) of adjoining cells use the same frequency band for transmitting their signals, inter-cell interference (ICI) occurs at user equipment (UE). To suppress ICI, UE needs to know the channel state information (CSI) between each of its surrounding BSs and itself. Therefore, multi-cell channel estimation is required. However, ICI also degrades the channel estimation performance. In this paper, we propose an iterative multicell minimum mean square error (MMSE) channel estimation scheme. To employ the MMSE principle, UE must know the covariance matrix of its estimating channels before conducting estimation. To cope with this contradiction, our proposed scheme estimates and updates the covariance matrix as well as the channels themselves through iterative processing. From computer simulation, we can show that the proposed scheme estimates the covariance matrix effectively, and achieves a higher channel estimation performance than the multi-cell least square (LS) channel estimation scheme.