Abstract: By using the TM images of Tianjin in 1999, 2003, and 2007 as the data source to acquire the wetland transition matrixes, and with CA-Markov model, this paper simulated the dynamic changes of wetland landscape pattern in Tianjin. The total area of wetland patch was increased from 3120.89 km² in 1999 to 3281.34 km² in 2007, then to 3297.43 km² in 2015, and would be decreased  to 3285.78 km² in 2023. The percentage of natural wetland would be decreased continuously from 33.0% to 30.2%, while that of constructed wetland would be increased to 69.9%. Twenty variables were selected and divided into two types for establishing a driving factors index system. The analysis of driving force for the wetland landscape changes showed that the first principal component had positive correlations with GDP, per capita GDP, total fiscal revenue, secondary industry percentage, total population, non-agricultural population, freight turnover quantity, passenger turnover quantity, and port throughput, and negative correlations with primary industry percentage, tertiary industry percentage, and cropland area.

Key words: plant community, TWINSPAN, environmental interpretation., DCCA