Autism Spectrum Disorder (ASD) is a one type of neuro-disorder. Due to this disease a person behaviour is being changed while they make interaction and communication with others. Autism Spectrum Disorder (ASD) is also called as “Behavioural Disease”. If a person is being effected from ASD, Symptoms are usually identified in the first two years of their life. According to so many researches and studies ASD problem is starts with childhood and continues to keep going throughout their life. Recent studies proven that ASD is gaining it’s momentum and growing faster than ever. So it makes somewhat expensive and time taken in finding out autism symptoms through screening tests. With the advancement of Deep Learning Techniques, we planned to work on ASD and propose a methodology in deep learning for earlier prediction of ASD. In this project work we are going to work with some of the Deep Learning Algorithms and also with some of the Advanced Machine Learning Techniques. Our project aim is to find out the best algorithm fit for early prediction of ASD based on performance metrics like Accuracy, Precision Score, Recall Score and F1 Score. For this project work we get the dataset from Kaggle repository, which consists of 21 attributes and 704 records. All considered algorithms have been implemented on python-programming using Spyder IDE with some important packages and libraries like pandas, numpy, matplotlib, sklearn and seaborns.