Abstract: With the advent of the post genome era, the number of protein sequences has grown rapidly. It is difficult to use experimental methods in largcscale analysis of protein subcellular location. In recent years, by using the various features of protein information (the encoding methods) , the algorithms that automatically predict protein subcellular location have been rapidly developed. In this paper, the encoding methods of protein sequence were surnrnarized, and the existing problems of encoding methods and the possible future research were discussed.