The common transition metal ions include Fe2+,Fe3+,Mg2+,Mn2+,Zn2+,Cu2+and so on.They play the role of stability,helping maintain protein structure and regulatory in life[1].Some proteins perform their function by interacting with other ligand,spicily transition metal ions.So the research of identification metal binding residues has the important meaning that guiding value for the development of molecular drug design.The based-sequence recognition of transition metal ion binding residues is an important way to understand this kind of molecular mechanism.Then we collected a non-redundant datasets of transition metal ion binding residues from the BioLip[2]database which have a resolution better than 3(A),sequence length > 50 residues and sequence identity below 30%.The datasets consists of 92,217,1461,459,1428 and 117 chains,382,1057,5212,2124,6408 and 485 biding residues for Fe2+,Fe3+,Mg2+,Mn2+,Zn2+ and Cu2+ respectively.