On 11 August， 2020， the National Meteorological Station in Lushan County， Sichuan Province， experienced an extremely abnormal precipitation event that far exceeded historical averages. The 24-hour rainfall reached 425.7 mm， ranking second in historical records for Sichuan Province. Within a 3-hour span， rainfall reached 348.6 mm， causing significant meteorological disasters. To accurately assess the impact of this rainfall event in the local history， annual maximum rainfall data spanning 42 years (1980-2021) with various durations (60，120， 180， 360， 720， 1 440 minutes) were collected from the Lushan National Meteorological Station. Employing a combination of chi-square test sorting， mean square error， residual sum of squares， and Wasserstein distance， as evaluation criteria， while considering the consistency of precipitation distributions across different durations， the optimal distribution， Johnsonsb distribution， were selected from among 80 probability density distribution functions. Reoccurrence period wasinferred by using the inverse function of the distribution. Results indicate that this method， when compared to commonly used limited distributions such as Pearson Type III， Gumbel， and Gamma distributions in hydrological and meteorological contexts， minimized overall mathematical fitting errors and considered engineering practicability. Particularly during anomalous precipitation events， the computed reoccurrence periods were more in line with actual scenarios. Through comparison across counties including Yucheng， Tianquan， Baoxing， and Shimian， the reoccurrence periods calculated by using this method effectively reflected precipitation reoccurrence characteristics.