It is crucial to conduct the land use/cover research to obtain the global change information.Urban area is one of the most sensitive areas in land use/cover change.Therefore land use/cover change in urban areas is very im-portant in global change.It is vital to incorporate the information of urban land use/cover change into the process of decision-making about urban area development.In this paper,a new urban change detection approach,urban dynamic monitoring based on objects,is introduced.This approach includes four steps:1)producing multi-scale objects from multi-temporal remotely sensed images with spectrum,texture and context information;2)extracting possible changed objects adopting object-oriented classification;3)obtaining shared objects as the basic units for urban change detection;4)determining the threshold to segment the changed objects from the possible changed objects using Otsu method.In this paper,the object-based approach was applied to detecting the urban expansion in Haidian District,Beijing,China with two Landsat Thematic Mapper(TM)data in 1997 and 2004.The results indicated that the overall accuracy was about 84.83%,and Kappa about 0.785.Compared with other conventional approaches,the object-based approach was advantageous in reducing the error accumulation of image classification of each datum and in independence to the radiometric correction and image registration accuracy. It is crucial to conduct the land use / cover research to obtain the global change information. Urban area is one of the most sensitive areas in land use / cover change. Wherefore land use / cover change in urban areas is very im-portant in global change.It is vital to incorporate the information of urban land use / cover change into the process of decision-making about urban area development.In this paper, a new urban change detection approach, urban dynamic monitoring based on objects, is introduced.This 1) producing multi-scale objects from multi-temporal remotely sensed images with spectrum, texture and context information; 2) extracting possibly changed objects adopting object-oriented classification; 3) obtaining shared objects as the basic units for urban change detection; 4) determining the threshold to segment the changed objects from the possible changed objects using Otsu method. In this paper, the object-based approach was applied to detecting the urban expansion in H aidian District, Beijing, China with two Landsat Thematic Mapper (TM) data in 1997 and 2004.The results indicated that the overall accuracy was about 84.83%, and Kappa about 0.785.Compared with other conventional approaches, the object-based approach was advantageous in reducing the error accumulation of image classification of each datum and in independence to the radiometric correction and image registration accuracy.