Autism Spectrum Disorder (ASD) is a sort of developmental issue of the nervous system, with center impedances in social connections, creative mind, communications, adaptability of thought,intrigue andrestricted range of activities. Examination of electroencephalographic (EEG) signals based on autism is explored in this work. Even so, it is critical to identify autism by the analysis of the EEG signal. Hence feature extraction based on the EEG signals takes part a prominent role in autism recognition. A practical feature extraction technique variational mode decomposition (VMD) to diagnose autism is narrated in this paper. Further, the features extracted are fed to classifiers ANN, KNNand SVM to stratify autism.SVM classifier shows a finer classification performance when compared to extant techniques.