Motor imagery (MI) based brain-computer interface (BCI) is a communication device that helps motor disabled patients to interact with the surrounding through their brain signals. But, it has low performance due to huge variations of brain patterns among the patients. The main reason behind is that the difference in spatial and temporal distribution of the brain signals. In order to boost the efficiency of the system, this paper combined features obtained from the Hilbert transform (HT) and second order difference plot (SODP). The proposed technique decomposed raw electroencephalogram (EEG) signals into multiple sub-bands with distinct frequency bands. The event-related patterns (ERPs) and MI features for each band were extracted through the HT and SODP. The obtained ERPs and MI features were fed into a multi-class support vector machine (SVM) for decoding brain activities. Two different benchmark datasets (BCI competition-III and IV) were used to evaluate the performance of the proposed method. The results show that the mean classification accuracy (%CA) and Cohen’s kappa coefficient (K) obtained from the proposed technique are higher than state-of-the-art techniques.