Abstract In the light of the wide divergence in the results of the researches into institutional investor and volatility of stock price in the existing literature,the authors hold that both herd behavior model and Fama MacBeth cross section regression method made the estimate of volatility of stock price biased because neither of them considered the correlation between the two. The authors combined herd behavior and Fama MacBeth cross section regression method to analyze the factors influencing the volatility of stock price. The empirical findings show that herd behavior is related to the ratio of institutional shares, scale and other variables. At the same time, after considering the variables of herd behavior and turnover rate, the coefficient of the influence of the ratio of institutional shares on the volatility of stock price becomes smaller obviously and different institutional investors have different influence the volatility of stock price, which proves the authors'corresponding proposition.