Aim and Background: The autism spectrum disorder is increasing in all parts of the world, so that in 2018, one in every 59 people with autism is reported. Given the importance of adaptive behavior in everyday life at home and at school, as well as social costs and care during their lives, the importance of early diagnosis in autism is highlighted, thereby improving cognitive, adaptive behavior and reducing the severity of autism.

Methods and Materials: The sample consists of 170 children, 100 autistic children and 70 healthy and juvenile children. Using a perception multi-layer neural network, a clinical decision support system was designed to predict autism based on signs and symptoms.

Findings: The average accuracy of the neural network design was 96.11% after performing 10, as a result of a carefully designed system; it could be a reliable assistant and supporter of this area for the diagnosis of autistic children.

Conclusions: The mean age of diagnosis of autism is 4 years and 4 months (52 months), which does not significantly differ with gender, race, and ethnicity. Therefore, the design of the clinical decision support system in this study for a range of 3 to 6 years old can be used as an important, necessary and reliable screening tool to prevent delay in early diagnosis and intervention.