The paper deals with two-dimensional (2D) channel estimation of Orthogonal Frequency Division Multiplexing (OFDM) system in slow fading wireless channel. We concentrate on two channel estimation schemes: Least Square (LS)+Weighted BiLinear (WBL) and LS+Linear Minimum Mean-Squared Error (LMMSE) where the first method is proposed in this paper. After theory analysis and simulation in Typical Urban (TU) channel, we find that LS+LMMSE achieves the optimal performance by exploiting prior knowledge of channel whereas LS+WBL, without requiring channel knowledge and with only half of the computational amount of LS+LMMSE, approaches LS+LMMSE in Bit Error Ratio (BER) performance when the distance of two adjoining pilot symbols along frequency direction is sufficiently small. This makes LS+WBL very suitable for wideband wireless applications.