Accurate weather forecasting remains a significant challenge due to the complexity of the Earth's atmospheric system and the chaotic nature of weather processes. Forecasting rainfall at specific locations is particularly difficult because traditional grid-based ensemble forecasts often fail to capture localized variations. The study focuses on probabilistic rainfall forecasts at station points using the EcPoint post-processing technique developed at ECMWF. The research examines how EcPoint enhances forecast reliability and discrimination skill across different lead times, seasons, and topographical variations in three provinces of China which include; Anhui, Zhejiang, and Jiangsu. it is part of broader efforts to refine ensemble-based precipitation forecasts, as raw ensemble outputs often contain systematic biases and struggle with point-scale accuracy, especially in complex terrain. By post-processing ensemble forecasts, EcPoint aims to correct these biases and provide more accurate localized forecasts. The findings of the study demonstrate that EcPoint significantly improves the skill and accurancy of ensemble rainfall forecasts at points Compared to raw ensemble forecasts.