The legal regulation of the training datasets of the intelligent assisted trial model is different from that of the ordinary training datasets, and the legal regulation of the model datasets has its own unique requirements based on the characteristics of the judiciary itself. Studying the regulatory requirements for training datasets in intelligence-assisted trial models will help enrich and improve the application of judicial trial theory in the field of AI-assisted justice. This paper, from the perspective of judicial practice and judicial value in various countries, analyzes the regulatory attitudes of each country towards the application of training data in the judicial field, and examines the requirements of the regulatory path from the perspective of judicial value. By using literature analysis, case analysis and comparative research methods, this paper attempts to explore the reasonable legal regulation path for the training datasets of intelligence-assisted trial models in the context of China.