The invention discloses a full-automatic segmentation method for kidney tumors. The full-automatic segmentation method comprises an experimental environment part and an algorithm research part. The experimental platform of the invention is a Windows 10 operating system, and the display card of the experimental platform is NVIDIA GeForce GTX 1080Ti. The algorithm research mainly comprises coarse segmentation of a CT image kidney and fine segmentation of a kidney tumor. The operation process of the whole system is as follows: 1, installing a Windows 10 operating system, and configuring an Anaconda version to be Anaconda 3, a python version to be 3.6, a CUDA version to be 10.0, a PyTorch version to be 1.6 and other required installation packages; 2, carrying out preprocessing operation on the data set, wherein the operation is operated in a pycharm which is provided with an Anaconda environment; 3, segmenting the processed data set by using the RAUNet. 4, carrying out post-processing on the pictures output after training. and 5, evaluating the model on the test set through an evaluation index. The main purpose of the invention is to promote the development of intelligent medical treatment, save labor cost, improve segmentation precision, and prevent diseases in advance, thereby ensuring human health.