Autism Spectrum Disorder (ASD) is a neuro-developmental condition that profoundly impacts the daily lives of individuals, affecting their behavior, social interactions, and communication abilities. The severity of the condition can be reduced by implementing early measures, taking into account the fact that the disease is difficult to completely eradicate. Due to the complexity and heterogeneity making an ASD diagnosis in a youngster might be difficult. Most individuals with ASD suffer with social behavior and communication, which among other things makes it difficult for them to learn. Early identification of ASD can lead to quicker interventions and better long-term results. There are several ways to differentiate ASD individuals from non-ASD people. However, other methods that can enhance ASD detection performance and classification must be developed. In a number of artificial intelligence tasks, the cutting edge of learning algorithms has lately been developed greatly by deep learning approaches. Object detection, speech recognition, and machine translation are examples of these artificial intelligence tasks. Hence, in this work, a survey is given on deep learning methods for ASD. In this survey, different research works on deep learning strategies for ASD identification and categorization are discussed. Finally, this work provides an overview on ASD detection techniques.