Abstract: This paper constructs the asset price bubble tests with heteroscedasticity White adjustment for the unit root data process where the error follows a generalized autoregressive conditional heteroskedastic process with order one. Theoretical research shows that the distribution of tests with heteroscedasticity White adjustment shares the same distribution with that of non adjusted in large samples. Simulation shows the heteroscedasticity White adjustment tests both has the lowest level of distortion and the satisfactory test power in large samples. Empirical research shows the necessity and rationality of using heteroscedasticity White adjustment to the test. The theoretical contribution of this paper is the derivation of the distribution for heteroscedasticity White adjustment test, which provides a new way for testing whether the asset price with heteroscedastic structure produced by disturbance has bubbles.