By employing the monthly average precipitation from 87 stations in Guangxi in June and NCEP/NCAR reanalysis data, the correlation between the interannual increment of monthly precipitation in Guangxi in June and the 500 hPa geopotential height field in the previous period from 1960 to 2021 is under investigation. Selecting the precursor signals that impact the precipitation anomaly in Guangxi in June occurs as part of this investigation. An ensemble forecasting model of the interannual increment of monthly precipitation, constructed by combining the fuzzy neural network and entropy metric method, is in continual operation. The cross-check of the prediction model from 1960 to 2013 and the independent sample test from 2014 to 2021 happen regularly. Results display a relative high prediction accuracy of the model, with a correlation coefficient of 0.93 between the predicted and actual values of the interannual increments of the return sample, passing the significance test of α=0.001. There is a return-year homogeneity rate of 87.5%, a fitted mean absolute error of 26.64 mm, and a fitted mean relative error of 9.06 %. This model is more stable than the prediction model built by the traditional stepwise regression method. For this reason, the entropy metric-fuzzy neural network ensemble prediction model sees better prospects for operational forecasting of short-term climate drought and flood trends.