By quality control methods (i.e., time series inspection, selection of stability data and eliminating the abnormal values), high-quality CO₂ datasets of Zhangjiagang Station in Suzhou from 2015 to 2018 were obtained to study the background concentration and source-sink characteristics of CO₂ in Suzhou. By Moving Average Filtering (MAF) method, CO₂ datasets have been separated as background and non-background data. The background data showed that the diurnal variation of CO₂ background concentration showed a single peak structure, and the peak and valley value was around 5:00 a.m. and 15:00 p.m, respectively. The seasonal variation showed a double peak structure, with the peaks in December and April. Both of the distribution of daily and seasonal variation of CO₂ were related to terrestrial ecosystems, meteorological conditions and human activities. From 2015 to 2018, the CO₂ concentration has increased. Particularly, the proportion of rising concentration has significantly increased, which may due to the human activities. On the contrast, the variation of the absorption concentration was stable, which inferred that the role of terrestrial ecosystem on CO₂ concentration was relatively stable. The source-sink analysis showed that, seasonally, the CO₂ uplifting concentration fluctuated slightly, while the absorption concentration showed remarkable seasonally variations. The diurnal variation of the rising concentration was a single-peak structure, with the peak value around 8:00 a.m., which may be traced to the high motor vehicle emissions in the morning. The absorption concentration was low in the daytime and high in the nighttime, which may be related to the plant photosynthesis and convective transport. We further analyzed the relationship between the CO₂ concentration and the wind field. The high CO₂ concentration has prevailed under the condition of calm wind in all seasons. Generally, in most wind directions, the concentration of CO₂ was negatively correlated with the wind speed. Particularly, the high CO₂ concentration was mostly accompanied at south to west-northwest winds. This may due to that the inland urban agglomeration was mainly located in southwest to northwest direction, and the construction area around the station was mainly located in southwest to north direction. Consequently, weak winds would lead to cumulative local emissions. However, the concentration of CO₂ was high at west-northwest winds. This may due to that the building area and inland urban agglomeration was in the west to north direction of the station. As the wind blew eastward, the concentration was low even when the wind is small, because the air from the sea is relatively clean.Moreover, the farmland and forest areas were mainly located in the east of the station, which was less affected by human activities. Our results have suggested that the concentration of CO₂ was not only related to the wind speed, but also related to the type of underlying surface and the long-distance environmental characteristics.