This paper elaborates on the principles of computer task scheduling, including goal setting, constraint analysis, and the working mechanism and process of traditional task scheduling algorithms. Analyzed the application of machine learning in task scheduling, covering aspects such as task feature extraction, task prediction, and resource allocation. At the same time, optimization strategies for computer automation task scheduling algorithms based on machine learning were proposed, including algorithm selection and improvement, parameter optimization and training method adjustment, and hybrid algorithm design, in order to provide theoretical support and practical guidance for improving the efficiency and performance of computer task scheduling.g.