The paper presents an adaptive channel estimation algorithm for orthogonal frequency division multiplexing (OFDM) systems in the presence of narrowband interference (NBI). The performance of the conventional linear minimum mean square error (MMSE) estimator degrades significantly in the presence of the narrowband interference. A new weighted recursive least M-estimate (RLM) algorithm is adopted for suppressing the adverse effect of time-varying narrowband interference to OFDM systems. Simulation results demonstrate the improved robustness of the proposed algorithm.