Precipitable water is an important intermediate parameter to estimate the probable maximum precipitation. In order to verify the accuracy of different methods for precipitable water estimation, radiosonde data and ground dew point temperature data at 13 stations in SE China coast region during 1998-2010 have been used to compare the precipitable water estimated by 3 methods, which are the specific humidity stratification integral, the surface dew point empirical formula method and the pseudo adiabatic calculation method, respectively. These stations are located in the five coastal provinces (region) of SE China and Hong Kong. The results show that the precipitable water calculated through the surface dew point method is the smallest, while the estimation through pseudo adiabatic method is the largest. The bias varies with latitude, the latitude is lower, the larger relative error. This bias is more obvious in the rainy days. The results of precipitable water by using ground method and pseudo adiabatic method both have strong correlations with the one based on specific humidity stratification integral method. The temporal variations of the atmo - spheric precipitation estimated by the three methods are consistent. This paper suggests that, in estimation of PMP in terms of moisture maximization, when the atmospheric precipitable water is calculated via pseudo adiabatic method based on the maximum 12-hour persisting surface dew point, a correction of latitude-induced difference should be taken into account.