The existing literature has confirmed that analyst bias can exacerbate stock price crash risk, but it is not clear how analyst bias affects stock price crash risk. Considering the impact of information asymmetry and investors’ own differences, this paper argues that the increase of analyst forecast bias will improve investors’ heterogeneous beliefs; under the condition of short selling restriction, the improvement of investors’ heterogeneous beliefs will boost stock price to a higher level, thus exacerbating stock price crash risk. Therefore, the impact of analyst forecast bias on stock price crash risk is likely to be mediated by investors’ heterogeneous beliefs. To test this theoretical prediction, this paper selects China’s A-share listed companies from 2008 to 2019 as research samples, constructs a mediating effect model, and makes an empirical analysis to study the role of investors’ heterogeneous beliefs in the impact of analyst forecast bias on stock price crash risk. The results show that:（1）Analyst forecast bias has a positive impact on investors’ heterogeneous beliefs. The larger the analyst forecast bias in the previous period, the greater the investors’ heterogeneous beliefs in the current period.（2）Investors’ heterogeneous beliefs can significantly enhance stock price crash risk. The increase of investors’ heterogeneous beliefs will exacerbate stock price crash risk.（3）Analyst forecast bias will improve investors’ heterogeneous beliefs, and then exacerbate stock price crash risk. Investors’ heterogeneous beliefs play a partial mediating role in the impact of analyst forecast bias on stock price crash risk. This conclusion basically validates the above theoretical analysis.

The contributions of this paper mainly include two aspects: First, it constructs a complete impact path and transmission mechanism. Although many scholars have studied the relationship between analyst bias and stock price crash risk, they have not further analyzed the detailed mechanism. They often put the investor behavior outside the research framework. In this paper, analysts, investors and stock prices of listed companies are included in the research framework at the same time. The impact path starts with analyst forecast bias, mediates investors’ heterogeneous beliefs, and ends with stock price crash risk. Second, it covers a number of market participants, and the research conclusion has important practical significance. Listed companies should strive to improve information transparency, decrease analyst bias, so as to reduce investors’ heterogeneous beliefs. It is one of the effective ways to avoid the stock price crash. At the same time, it also has important reference significance for securities industry practitioners, investors, regulatory authorities and the government to effectively realize and prevent stock price crash risk.