Aiming at the problem of automatic lung CT segmentation，an improved U-Net model is proposed．The model fully utilizes the advantages of U-Net model and Ｒecurrent Convolutional Neural Network( ＲCNN) ，andit helps extract very low-level features and ensure better performance with fewer numbers of network parameters．Firstly，in the encoding and decoding units，a recurrent convolution layer replaced the forward convolutional layersof the original U-Net models． Secondly，an effective feature accumulation method was utilized from one part of thenetwork to the other． Finally，the cropping and copying units were removed from the original U-Net model and onlythe concatenation operations were retained． To evaluate the performance of the proposed scheme，the comparativesegmentation experiments with state-of-the-art models were conducted in the LUNA16 datasets． The experimentalresults proved a superiority of the proposed method on segmentation accuracy．