In this paper, we propose a new channel estimation algorithm for Orthogonal Frequency Division Multiplexing (OFDM) systems, which is extended to the Alamouti Coded OFDM case by extracting required equations. At the first stage of this algorithm, the channel response is estimated for pilot sub-carriers and then, the channel response for data sub-carriers is interpolated by Gaussian Radial Basis Function Network as it is an efficient nonlinear interpolator. The post-processing and filtering on the estimated channel taps energy are proposed to improve the estimation performance and reduce the noise effects of primary estimation, which is caused by LS estimation at pilot data. At next step, the accuracy of our channel estimation algorithm is improved by using iterative decision feedback. The decision feedback method is used to update channel parameters in each OFDM symbol. The effectiveness of proposed techniques is demonstrated through the simulation of OFDM and an Alamouti coded OFDM system with two-transmit and two-receive antennas. Finally, the results are analyzed and compared with previous conventional estimation algorithms. Simulation results show that the proposed algorithms achieve efficient performance in many practical situations.