The study of the optimal sampling frequency is very important in high-frequency financial data.  The study of optimal sampling frequency home and abroad was introduced and commented.  Considering the characters of highfrequency data in Chinese stock market, a more succinct and convenient method of the optimal sampling frequency was provided and the theorems of the bias of realized bipower variation and weighted realized volatility were proved.   The results of the empirical study of Shenzhen stock market are also given.