Through rapid innovation in the health care sector, several types of biomedical signals, including the electrocardiogram (ECG), electroencephalogram (EEG), positron emission tomography (PET), electromyogram (EMG) and electrogastrogram (EGG), now play an important role in the accurate analysis and detection of many health problems, such as heart related problems, sleep disorders, brain abnormalities etc. In the current chapter, an efficient approach based on the Fourier decomposition method (FDM) is proposed for noise removal and classification of epileptic and healthy EEG signals. This work is composed of three main steps. In the first, preprocessing of the EEG signal is performed to remove noise or artifacts present in EEG recordings. In the second step, the FDM is utilized to fragment non-stationary data into a set of orthogonal basis functions known as Fourier intrinsic band functions (FIBFs). In the third step, FIBF based features are extracted and fed to various models for the accurate classification of EEG signals. The simulation results reveal that the current study has achieved superior performance compared to the existing techniques discussed in the literature. © IOP Publishing Ltd 2020.