The widely use of automatic weather station has greatly improved the accuracy of meteorological monitoring, providing important support for forecasting and warning, meteorological services, climate analysis and scientific research. Based on the daily precipitation data of the national and automatic weather station in 2015-2021, the multi-scale spatial and temporal distribution of precipitation, the spatial category and intensity characteristics of heavy rainfall, and a brief weather background were analyzed. The main conclusions are as follows:(1) Based on the area-density weights of the rainstorm stations in each district, the spatial classification index of the heavy rainfall is developed with 4 categories of local, regional, wide-range and territory-coverage considering the independence of the assessment factors and the non-uniform distribution of the stations. Based on the area weight of the number of rainstorm stations and the disaster-causing effects of different magnitude rainfall, the assessment index of heavy rainfall intensity is developed. By using the percentile method to determine the intensity grade of heavy rainfall, the spatial range and intensity of heavy rainfall are closely connected.(2) The precipitation in Xiamen was characterized by significant regional heterogeneity and local characteristics due to the topographic and geomorphological characteristics of mountains, sea and bays. The mean annual precipitation and heavy rainfall frequency increased gradually from the coast to inland, and the frequency of rainstorm was closely related to the distribution of topography. (3) In 2015-2021, there were 106, 37, 16 and 5 rainstorms of the aforementioned category in Xiamen, accounting for 65%, 22%, 10% and 3%, respectively.Heavy rainfall occurred in every month, and concentrated in the main flood season from May to September, with peaking in June and August.There were 8 extremely heavy, 24 heavy, 33 relatively and 99 general cases, and the total intensity and average intensity ranked firstly in 2016 and the weakest in 2020. The spatial category of the 8 extremely heavy events was territory-wide scale, while the 99 general heavy rainfall were all local-range events.(4) The main influencing systems of the regional torrential rain processes included warm-cold air activities, tropical cyclones, southerly stream, northward uplift of convergence zone, low-level wind shear and strong convection, etc. The wide-range especially the territory-wide heavy rainfall caused mainly by warm-cold air activities and tropical cyclones. A rare territory-coverage winter heavy rainfall event occurred on 9_th^{Dec 2015 due to the impacting of} weak cold air and the strong warm moist airflow.The results can be used as a scientific reference for monitoring, forecasting, assessment and service of heavy rainfall.