Based on the land use/cover maps of 1990, 2000, and 2010, topographic factors, and geographic elements, a CA-Markov model consisting of Markov transition matrix, multi-criteria evaluation, and cellular automata was developed to simulate the change trends of the future land use and landscape patterns of Dalian, Liaoning Province. The future land use pattern of Dalian was optimally allocated by the method of fuzzy multi-objective programming, based on the characters of land use structure, society, economy, and natural environment. The results indicated that in 1990-2010, the rapid development of Dalian showed the characteristics of the continued expansion of urban area and the reduction of cropland and woodland area. With the present speed of urban development, the landscape pattern and land use cover would have a great change, and the landscape fragmentation would be exacerbated. To optimize the land use structure could meet the demand of the future sustainable development of Dalian.