Abstract: Accurate multi-organ segmentation in medical images holds significant importance for various clinical application and pharmaceutical development.However, traditional image processing methods based on manual feature design struggle with the complex textures and shapes in medical images.In recent years, with the rise of artificial intelligence, end-to-end deep learning approaches have demonstrated powerful potential in automated medical image analysis.Particularly, U-Net series of networks based on convolutional neural networks and Transformers have achieved precise semantic segmentation of medical data, which has significantly enhanced the accuracy of medical diagnosis and treatment in clinical decision-making and efficacy assessment.This paper reviews current deep learning-based algorithms for organ segmentation in medical images, focusing on the development of U-Net family networks and the application of multiorgan segmentation in medical advancements.