Abstract: Taking Baicheng， the western city in Jilin Province as study area， the present study used 2000， 2010 and 2014 TM/LOI remote sensing images and applied human\|computer interaction interpreting technology to extract land use information. The shifting area between different land use types were determined from 2000 to 2010 by spatial and statistical analysis， and the land use situation in 2020 was also predicted by the CA\|Markov model based on the three series remote sensing images. Referring to the revised Chinese terrestrial ecosystems per unit area service value table， the future ecosystem service value of the study area was estimated. It was shown that the land use types will undergo significant changes by 2020 in Baicheng. The area of dry land， paddy fields， woodland， grassland and construction land increased， while， the wetlands and unused land area exhibited a decreasing trend. There was no significant change in the value of ecosystem services from 2010 to 2020. It was slightly reduced from 5021×108 yuan in 2010 to 4919×108 yuan in 2020. The accuracy testing and sensitivity analysis verified the reliability of the present study. Thus， the present study would not only provide new methods for the forecast of ecosystem services value， but also have great significance to the regional sustainable development.

Key words: ecosystem services value, forecast, CA\, Markov, Baicheng City