Dynamic transboundary water resources allocation based on inflow prediction results is an important task for water resources management in river basins. This paper takes the watershed management agency as the leader and the associated area as the follower, and proposes a two-level asymmetric Nash-Harsanyi Leader-Follower game model considering inflow forecasting errors. In the proposed model, the Monte Carlo method is used to analyze the uncertainty of various stakeholder allocation results and the response regularity to the total water resource uncertainty. The quantitative relationship between the allocation results of stakeholders and the mean and standard deviation of total water resource uncertainty is subsequently established. The Huaihe River basin in China is selected as a case study. The results show the following: (1) the water allocated to the watershed management agency and three provinces has a normal distribution when the inflow forecasting error obeys the normal distribution; (2) the sum of the mean of the water allocated to stakeholders equals the mean of the forecast water resource and the sum of the standard deviations of the water allocated to stakeholders equals the standard deviation of the forecast water resource; (3) the mean and standard deviation of the allocation results have a good linear relationship with the mean and standard deviation of forecast water resource; (4) the distribution parameters of the stakeholder allocation results can be directly derived from the distribution parameters of the forecast information, thus aiding the stakeholders in making decisions and improving the practical value of the method.