In this paper, noise power and signal-to-noise ratio (SNR) estimations for OFDM based wireless communication systems are studied. Noise power estimation, which takes into account the color and variation of noise statistics over OFDM sub-carriers into account, is considered. Instead of averaging the instantaneous noise samples estimates over all of the OFDM sub-carriers, dividing the total number of sub-carriers into several sub-groups and averaging the sub-carriers separately within each sub-group is proposed. Also, these averaged estimates over each sub-group are further averaged across several OFDM symbols in time to enhance the performance of the noise power estimates. The proposed solution provides many local estimates, allowing tracking of the variation of the noise statistics across OFDM sub-carriers, which is particularly of use in subband adaptive modulation OFDM systems. The performance of the proposed method is evaluated via computer simulations. It is observed that the proposed solution can estimate local statistics of the noise power when the noise is colored. When the noise is white, the proposed algorithm works as good as the conventional noise power estimation schemes, showing the robustness of the proposed method.