Introduction: Autism Spectrum Disorder (ASD) is a deficit in brain development. It is characterized by a wide spectrum of conditions, such as challenging social skills, repetitive behaviours and difficulty in speech. At present, diagnosing children with ASD at a younger age is challenging. Currently, visual analysis is practised at hospitals which may lead to misinterpretation hence a quantitative analysis is required for early detection and intervention. Objectives: This research analyzes electroencephalography (EEG) signals of children with ASD and typically developing (TD) and to attain a better classification accuracy in identification. Materials and Methods: 10 ASD and TD children were considered for the study. Children were made to sit in front of a visual screen and asked to watch a video for ten minutes. During this period EEG signals were acquired, to analyze the difference in characteristics between ASD and TD children. Results: EEG signals were acquired from 19 channels. They are preprocessed and Recurrence Quantification Analysis (RQA) is applied to the resulting signal. The features extracted are then fed to different types of SVM classifiers. The responsive brain regions were identified and their contribution to RQA features was analyzed. Conclusion: Responsive channels were identified as Fp1, Fp2, F3, F4, Fz, Cz, O1, O2, T3 and T5. The features extracted from these channels were fed to SVM classifiers out of which quadratic SVM presented an accuracy of 81.8%. In future, larger data sets must be considered for validation and different RQA parameters must be considered for better accuracy.