Abstract: The aim of sign language recognition research is to let the computer to understand what a deaf 'said' using the sign language so that the computer can output the recognition results in the form of text and/or speaking language and help us to communicate with the deaf. The characteristics of Chinese Sign Language (CSL) are analyzed in the view of sign language recognition. The analysis and comparison of the two existing sign language recognition methods, the machine-vision-based method and the body-instrumentation-based method, are also done and a new clarification method for the CSL is also proposed.