TRMM daily precipitation data were validated compared to thirteen precipitation observation station data in Irrawaddy basin, and a calibration method of TRMM 3B42 data based on water balance concept was developed. SWAT model was used to calculate water balance, and then the TRMM data calibration ratio was inferred based on it. Average slope of sub-basin was the independent variable, and calibration ration was the dependent variable for the regression analysis. The calibration ratio of sub-basins without hydrological observation stations were calculated according to the average slope, and then the TRMM precipitation were calibrated according to the ratio in the whole basin. Results showed that TRMM annual and monthly precipitation had a highly correlation with observed data, but had a bad bias. Annual precipitation of TRMM precipitation after calibration were nearly the same as observed; monthly precipitation curve were similar with observed. Daily precipitation forecast rank was improved from medium to well, and there were 2 station forecast bad, 7 forecast well and 4 forecast medium before calibration, but 6 forecast well, 2 forecast medium and 5 forecast excellent after calibration. The results proved that TRMM data precision were highly improved over different temporal scales in Irrawaddy basin through this method, data for hydrological and water resources analysis were also provided.