Chinese Phonetic-Character Conversion（CPCC） is an important issue in speech recognition and Chinese sentence keyboard input system.The appoaches based on large amount of corpus statisties Markov models become more and more popular today

The CPCC based on Chinese character N-gram （C-CPCC） has the advantage Of having a smaller statistics data library and simple algorithm. but has the drawback of lower accuracy of conversion

while that based on Chinese word N-gram（W-CPCC） is on the contrary.This paper presents a word-self-made CPCC algorithm based on the Chinese Character Bigram. which not only has foe C-CPCC’s advantage of having a smaller statistics data library

but also can take advantage of the W-CPCC. The experiment shows it can be easily realized with a higher accuracy of conversion.