More than 20 sequences of phospholipid hydroperoxide glutathione peroxidase (PHGPX) from a sequence database were analyzed. The analyses show that the primary structures of most PHGPX proteins have three highly conserved regions forming a catalytic center and have more than 50% amino acid sequence identity in common. However, two PHGPXs from bovine and swine with the same function have very low similarity with typical PHGPXs and do not have the three highly conserved regions. Thus, the PHGPX proteins are divided into two types: those with the three highly conserved regions, designated as PHGPX-I, and the others as PHGPX-II. In general, type I proteins are composed of ca.170 amino acid residues; a few of them have an extra signal peptide sequence at the N-terminal of the protein. The composition of plant and animal PHGPX amino acids is very different, with most plant PHGPXs being weak acidic, while most animal ones are alkaline. Another specific conservative motif is also found in plant PHGPX proteins. System evolution analysis shows that ortholog and paralog evolution models both exist in PHGPXs, with the plant PHGPX and the animal PHGPX diverging exclusively into two branches in PHGPX-I. The information revealed by the evolution tree agrees with the general species evolution process from low to advanced and from simple to complicated.