Autism Spectrum Disorder (ASD) is a neurodevelopment syndrome that has a great impact on the person's behavior, mental health and also the learning skills. It limits the ability of the affected personality to use verbal, social, and cognitive skills, among other abilities, and its symptoms are differed from person to person. Diagnosing ASD is very difficult process because it does not have any medical test like blood test to predict the symptoms of the ASD. To overcome difficulties of diagnosing ASD, various machine learning methods are used to increase the performance and accuracy rate, so that diagnosing the ASD can be predicted at the early stages. As a result, it can be utilized to make decisions in a situation where there is a lot of uncertainty. In this research, machine learning techniques evaluate their performance on an ASD dataset. As the results, Random Forest Classifier and Decision Tree Classifier got the highest accuracy with 100% and Support Vector Machine (SVM) and Naive Bayes algorithms got the accuracy of 97%.