Abstract:Analysis and prediction of land use pattern is significant not only to balanced economic development and ecological protection, but also to ecological sustainable use and management of land protection. The study obtained the land-use map on the bases of remote sensing images in 2005 and 2015 of Dongxiang County, Jiangxi Province of East China and used the map to acquire conversion data and landscape indexes of land use types by ArcGIS 9.3 and Fragstats 4.2. Based on the conversion data and landscape indexes of land use types, the landscape pattern of Dongxiang County from 2005 to 2015 was analyzed. CA-Markov model was used to simulate the dynamics of land use change with time and state, and to predict land use pattern in the next 10 years. The results indicated: Construction land and arable land increased by 71.82 km2 and 10.76 km2 respectively for the past 10 years, but water area, forest land and unused land decreased with an order of unused land >water area>forest land. Compared to the 2005, landscape patches distributed more evenly, landscape diversity in 2015 increased; Landscape shape tended to be simple and landscape patch connectivity became lower in 2015, which indicated the associations between landscape patches weakened in 2015. Human economic activity and land use policies have more obvious effects on landscape pattern. According to the forecast by CA-Markov model, it is predicted that by 2025 construction land will be increased by 68.72 km2, water area, forest land and unused land will continues to be decreased, but the arable land will be increased. The number of landscape patches and the landscape heterogeneity will increase, the patch shape will become a little more complex. landscape pattern will become more diversified and homogenized. In conclusion, landscape pattern of Dongxiang is gradually become more stable, homogeneous and diverse, but great efforts and reasonable plans are needed to protect arable land, to control the use of construction land, and to maintain the balance of the ecological environment.