Biomedical systems are inherently characterized by complex dynamic behaviour, reflecting their intricate and multifaceted nature. This complexity has made them a focal point for advanced research and analysis, particularly through nonlinear dynamic methods. In this paper, identification of electroencephalography (EEG) is carried through improved broad learning system (IBLS). The improved nature of the BLS attained by modifying the computation of weights. In IBLS, weights are computed using elastic net regression instead of ridge regression. This modification enhances both accuracy and efficiency. The comparative analysis with other prevalent approaches highlights the proposed method's exceptional efficacy and reliability.