Abstract: The European Centre for Medium-Range Weather Forecasts (ECMWF) model is a prominent numerical model extensively utilized in operational meteorological forecasting. Using ECMWF model precipitation forecast data and observed data for Chongqing from 2017 to 2019, a comprehensive analysis and evaluation of the model's forecasting capability was conducted using TS score, ETS score, false forecast rate, missing forecast rate, and BIAS. Our findings indicate a pronounced decline in TS, ETS, and BIAS as precipitation intensity and forecast lead time escalate. Conversely, the false forecast rate and missing forecast rate demonstrate a significant increase. The annual variation characteristics of the TS and ETS scores are significantly different for different precipitation levels. Except for heavy rainfall and above levels, the scores of other precipitation levels are lowest in August, and the forecasting capability is relatively better in autumn. In terms of seasonal variation, there is a noticeable BIAS in the forecasted precipitation area and amount, with the largest BIAS in winter and the smallest in autumn. Among different forecast lead times, the 24-h forecast performs the best. The distribution of the annual average daily precipitation shows a dry BIAS in the forecast, particularly pronounced in the western and northeastern parts of Chongqing. Analysis of the precipitation processes during the rainy season reveals that the model can effectively capture the different stages of precipitation processes, with the 24-h forecast being the closest to the observed precipitation.