Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition characterized by persistent impairments in social interaction, communication, and behavioural patterns. Early and accurate diagnosis of ASD is crucial for effective intervention and improved developmental outcomes. In recent years, machine learning (ML) techniques have been increasingly employed to support ASD diagnosis using heterogeneous medical data sources such as clinicalassessments,behaviouralquestionnaires, neuroimaging, genetic information, and speech video data. However, the performance of individual ML models is often limited by data scarcity, class imbalance, noise, and weak inter-modal correlations. Ensemble machine learning approaches, which integrate multiple predictive models, have emerged as a promising solution to improve robustness, accuracy, and generalization. This survey presents a comprehensive and systematic review of ensemble-based ML techniques for ASD classification using multimodal medical data. The paper analyses data modalities, feature extraction methods, ensemble architectures, fusion strategies, evaluation metrics, and validation techniques. Key challenges, research gaps, ethical concerns, and future research directions are discussed, with an emphasis on explainable and clinically deployable ensemble frameworks