 The study proposes an efficient change rule with transferability to detect both binary and multiclass changes in remote sensing data using an improved long short-term memory, LSTM, model. The proposed method utilizes a core memory cell to learn the change rule from the information concerning binary or multiclass changes, and three gates to control the input, output, and update of the LSTM model for optimization. The learn rule can be applied to detect changes and transfer the change rule from one learned image to another new target multi-temporal image without any extra learning process. The study demonstrates the superiority of the proposed method through binary experiments, transfer experiments, and multiclass change experiments, and makes three contributions. 1. The proposed method can learn an effective change rule for multi-temporal images. 2. The learn change rule has good transferability for detecting changes in new target images without any extra learning process. And, 3. Deep learning in recurrent neural networks is exploited for change detection under the framework of the proposed method. This article was authored by Habu Lu, Huilu, and Li Chaomao. We are article.tv, links in the description below.