Traditional approaches for detecting the changes in remote sensing images are mainly based on spectral information of the original images without utilizing other derived features thus may result in low detection accuracy and low integrity.To solve this problem we proposed an algorithm for detecting the changes of remote sensing images based on object-oriented idea and adaptive multi-features fusion.First the eCognition was used to realize image segmentation and extract the spectral textural feature and space characteristics.A neural network was built up to realize adaptive multi-features fusion and the final detection results were obtained.The experimental results show that adaptive multi-feature fusion method can reduce the probability of missing detection and false detection and improve the accuracy and integrity of the detection result.