A novel comprehensive classification system is proposed for recognizing the orthogonal frequency division multiplexing (OFDM) signal and extracting its parameters. An empirical distribution function -based Gaussianity test technique is first applied to distinguish OFDM from single carrier modulations. Cyclostationarity and correlation tests are then applied to estimate OFDM symbol duration and cyclic prefix duration. A bank of fast Fourier transform (FFT) bank filter is devised to detect the number of subearriers. Numerical results are presented to verify the efficiency of the proposed scheme.