Abstract: A forecast model of short-time heavy precipitation (greater than 20 mm h-1) for 1–12 h at 1 h intervals in Fujian is established based neighborhood with optimal Threat Score by real-time precipitation observation data from ground weather stations, 0–120 min Quantitative Precipitation Forecast from Sever Weather Automatic Nowcasting (SWAN-QPF) of China Meteorological Administration, and forecast precipitation data from global and regional models from April to September during 2021–2023. Neighborhood test is adopted, and the test radius is 40 km. The results show that the accuracy of nowcast in the couple of hours is greatly improved by using real-time precipitation observation to forecast the short-time heavy precipitation (Persistence forecast) compared with using multi-model optimal weight integration. After the optimal elimination threshold correction, the TS in the future first hour can reach 37.5% for 2021 and 32.2% for 2022 when the forecast precipitation is calculated with 10-min real-time precipitation before the forecast production, the neighborhood radius (Ri) of 0.5°, and the average of the top 5 heavy precipitation stations (Ntop=5) in Ri. And at the second hour, the TS can reach 22.2% for 2021 and 19.5% for 2022. The TS of the consensus forecast combining global and regional models with optimized weights can reach 16.2% for 2021 and 16.6% for 2022 at the first hour (18.0% and 14.2% at the second hour) when the forecast precipitation is calculated with the Ri of 0.6° for each model and Ntop=15. The SWAN-QPF short-time heavy precipitation prediction revised by the optimal elimination threshold (the Ri of 0.3° and Ntop=15) is also better than multi-model optimal weight integration, but less than using real-time observational precipitation in the first couple of hours. For the 3–12 h forecast, multi-model optimal weight integration is better than the other two. The above-mentioned multi-source data revised by the optimal elimination threshold are further integrated with optimized weights in the future 1–4 h, and multi-model optimal weight integration is adopted in the future 5–12 h to establish the forecast model of short-time heavy precipitation for 1–12 h at 1 h intervals. The parameters trained with 2021 and 2022 data are applied to forecast short-time heavy rainfall in Fujian in 2023, and the TS are 42.7%, 28.8%, 23.1% and 20.2% in the future 1–4 h, and all above 17% in the future 5–12 h.