Autism Spectrum Disorder (ASD) is a prominent form of neurological disorder that impairs a person's ability to engage socially and interact effectively with others. ASD is also characterised by a tendency toward repetitive and restricted behaviours. The costs associated with autism can rise during the time it takes to get a diagnosis and begin treatment, but many of these expenditures are long-term and will remain with a person throughout their lives. Optimization and machine learning methods have been used in many different industries and professions in an effort to improve results.In this paper, an bioinspired optimization-based classifier namely Adaptive Whale Optimization based Support Vector Machine (AWOSVM) is proposed for precisely detect ASD. AWO-SVM performs classification after optimization phase gets complete. AWO-SVM involves three different phases namely Exploitation Phase, Exploration Phase, and Classification Phase. Each phase of AWO-SVM plays a major role to predict ASD more accurately. The metrics "accuracy" and "F-Measure" are used to evaluate AWO-SVM on three distinct ASD screening datasets. When AWO-SVM results are compared to those of other classifiers, it is clear that AWO-SVM is superior in terms of accuracy in predicting ASD.