The social and relational abilities of an individual with autism spectrum disorder (ASD) are impeded until the end of their lives. In this paper, we endeavor to explore the practicality of utilizing Nave Bayes, Backing Vector Machine, Strategic Relapse, KNN, Brain Organization, and Convolutional Brain Organization for foreseeing and dissecting ASD. There are 292 cases and 21 qualities in the first and the second dataset for screening grown-ups for autism spectrum disorder, and it comprises of 704 cases with 21 qualities. The juvenile autism spectrum disorder (ASD) screening dataset has 21 qualities and 104 cases. Information for grown-ups, youngsters, and teenagers were evaluated for autism utilizing an assortment of machine learning strategies, and the outcomes emphatically recommend that CNN-based expectation models work best. Their precision was 99.53%, 98.30%, and 96.88%, separately