Abstract: This research aimed to explore the application of the combination of region growth and U-Net in the retinal vessel segmentation system, retinal vessels were automatically segmented through the developed system. This research provides doctors with accurate information on changes in the retinal blood vessel structure, which will improve the efficiency of diagnosis and treatment. Methods: Combined with U-Net network and region growth, the pre-processed retinal blood vessels were automatically segmented, and the algorithms were integrated into the retinal blood vessel segmentation system through the design of controls. Results: The average values of blood vessel segmentation performance indexes— accuracy, sensitivity, and specificity were 0.977 7, 0.768 4, and 0.982 1, respectively, and regional iterative growth could improve the segmentation effect of fine retinal blood vessels. Conclusion The system has the characteristics of simple interface and convenient operation.It realizes automatic retinal blood vessel segmentation with high precision and visualization, provides an effective application platform for doctors to observe the changes of retinal vascular structure, and also provides a thinking direction for doctors to judge the nature of lesions.