This study was conducted to develop the predictive model for species richness of breeding birds by multivariate analysis of land covers (percentage value of each cover in 300 $ imes$ 300m grid) including water area (WA), urbanized area (UA), green area (GA), forest area (FA) and agricultural area (AA) at Seongnam City in Korea. Fifty-nine species of birds were mapped on 155 grids in size of 300 $ imes$ 300 m from 2000 to 2001. Species richness of breeding birds was significantly regressed ($R^2=0.85$, n=155) by estimates of WA ($4.59{pm}1.13$, P<0.0001), UA ($6.50{pm}0.86$, P<0.0001), GA ($7.11{pm}1.25$, P<0.0001), FA ($8.51{pm}0.62$, P<0.0001), and AA ($4.59{pm}1.13$, P<0.0001). In validation of model, the results show no significant difference between predicted value of species richness and observed one. Developed model can be used as a predictive model of species richness of birds for selecting the proper location of corridors and parks in urban area.