We use the physical and chemical characteristics of the protein to combine with the tree model to predict the secondary structure based on the primary structure of the protein.