In this study, we propose RLA-CNN, a novel method of motion intention recognition, based on residual learning attention network. It combines the residual block and the learning attention module for motion intention classification on the EEG Motion Imagery Dataset. The average validation accuracy of 75.85%, 84.81%, and 84.69% is obtained, which outperforms the state-of-the-art by 5.65%, 10.06%, and 7.41%. This paper also gives the visualized analysis on the learned attention weights of RLA-CNN to explain the results.