Abstract: The quantitative three-dimensional cloud data is important in nowcasting and the modeling of weather and climate. Therefore, 5 schemes are designed to construct three-dimensional cloud amount data from FY-2C satellite data, radar data, ground observation data using LAPS (Local Analysis Prediction System) developed by NOAA ERSL. The roles of each data in LAPS system are also analyzed. Scheme 1 uses background data only, and Scheme 2 adds ground observation data. Scheme 3 employs background data and FY-2C satellite data, Scheme 4 uses background data and radar data, and Scheme 5 takes background data, ground observation data, radar data and FY-2C satellite data into consideration.The analysis indicates that every data is important in order to get more objective three-dimensional cloud distribution. Ground observation data gives information of cloud base and cloud amount for the lower atmosphere. Satellite infrared brightness temperature and visible reflectance provide cloud top height and cloud amount in the upper atmosphere. Radar data can help to construct three-dimensional cloud field in the middle and lower level. Combining all these data can provide more objective information of three-dimensional cloud amount.Comparing column cloud amount deduced by LAPS with satellite visible and infrared image shows that the cloud distribution when assimilating all these data is more consistent with real situation. Moreover, the satellite data is one of the most important data in cloud analysis in LAPS.