Abstract: Optical astronimical data are very important source of information. A significant scientific challenge is to separate stars and galaxies from the photometric data. Perform such classification is the goal of this article. Decision tree i sthe technique employed in this paper. Data from the Sloan Digital Sky Survey project were used for training and testing the scheme. The final sample is composed by 43289 stars and 452400 galaxies, where relevant photometric parameters were selected to identify stars and galaxies. The method presented good results.