The estimated rainfall from satellite is very important for engineering to management
planning and design. The technique uses the Global Satellite Mapping Precipitation (GSMaP)
product composed of precipitation retrievals from the Tropical Rainfall Measuring Mission
(TRMM) satellite and other polar-orbiting satellites to interpolates them with cloud motion
vectors derived from infrared images from geostationary satellites then the produce is a high
resolution dataset. Four other satellite-based datasets are also evaluated concurrently with
GSMaP. An hourly data from GSMaP_MVK and GSMaP_NRT in 2009 used in the
experiments, and also selected 116 rain gauge stations over Thailand to verify. The results
showed that GSMaP_MVK is a good the relation to rain gauge observations than
GSMaP_NRT in the whole area with the correlation coefficiencies (CC) is the average 0.80
that is positive values then it also indicates a strong positive linear relationship via a firm
linear rule, the standard deviation (STD) is the average 0.45 indicates that the data points tend
to very close, and the mean absolute error (MAE) measures the average magnitude of the
errors in a set to forecast error in the time series is the average is 2.49. Future work would
enable this to the multiple regression method to the estimated rainfall by satellite over the
whole Thailand.