On Sunday, 21/7/2024, the master’s thesis of the student (Mazen Radhi Sawadi), majoring in (Computer Engineering), was discussed in the Control and Systems Engineering Department in the Discussion Room (No. 9). The thesis title is: “Intelligent System for Detecting Autism Spectral Disorder and Allocating Suitable Educational program”. The discussion committee consisted of: 1 - Prof. Dr. Ashwaq Talib Hashem / Chairman 2- Asst. Prof. Dr. Mona Hadi Saleh / Member 3- Asst. Prof. Dr. Ahmed Mudher Hassan / Member 4- Prof. Dr. Muayad Sadiq Crook / Member and Supervisor Autism Spectrum Disorder (ASD) refers to a group of neurodevelopmental disorders, including autism, Asperger's Syndrome (AS), and Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS). The new diagnostic criteria for ASD focus on two core domains: social communication impairment and restricted interests/repetitive behaviors. There are three levels of autism spectrum disorder (ASD), which are described in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) and Childhood Autism Rating Scale, Second Edition (CARS-2). The three levels can be used as a springboard to understand better that autistic children have a variety of required support. ASD child's life can be significantly improved by early diagnosis of the disorder and intervention, depending on the necessary support. In this thesis, a mobile application for autism detection and level identification based on the Deep Learning model is proposed. It includes two stages: classification and level identification. The first stage classifies the child as either an ASD child or potentially normal, while the second stage identifies the ASD level based on traditional examination methods DSM-5 and CARS-2. This level is important in allocating the suitable education program that should be followed with ASD children. TensorFlow and Keras libraries are used for the suggested Deep Learning Convolutional Neural Network (CNN)-based model to perform feature extraction and picture categorization. The model is trained and tested using an ASD dataset from the Kaggle repository. The dataset used to train and test this model consisted of 2,122 excluded from the original dataset of 2,940 images due to the quality and unbalanced race. An accuracy of 97.3% and an Area Under Curve (AUC) of 99.8% are obtained from the suggested model. This application proves its high classification accuracy, ease of use, and fast decision-making capability, and this can surely help caregivers get an idea of standard learning strategy assignments by following with their child. On this occasion, we congratulate the student (Mazen Radi Sawadi) and wish him continued success.