According to the precipitation and δ~(18)O obtained during GAME-Tibet IOP, based on the knowledge that the δ~(18)O in precipitation is lower which is from the ocean air mass and higher from that of the local evaporation, the water vapor sources could be identified from the δ~(18)O value in precipitation. We attempt to give the identification criterion of δ~(18)O value. If we use δ~(18)O -20‰ as the criterion that the precipitation is formed directly from the ocean air mass, the contribution of ocean air mass may be not underestimated. And if we use δ~(18)O-13‰ as the criterion that the precipitation is formed directly from the local evaporation, the contribution of local evaporation may be not overestimated. According to such criterion, the proportion of the local-evaporation-formed-precipitation and the ocean-air-mass-formed-precipitation in total precipitation was estimated. The precipitation samples at site NODA, Anduo and AQB basically recorded the precipitation processes in June to early September. On the average of these three sites, the precipitation is 249.76mm. The precipitation formed directly from ocean air mass vapor is not beyond 80.08mm and the precipitation formed directly from local evaporation vapor at least is 117.05mm. That is to say, the proportion of the ocean-air-mass-vapor-formed-precipitation in total precipitation is not beyond 32.06% and the local-evaporation-vapor-formed-precipitation at least is 46.86%. The other 21.8% may come from the transportation of the monsoon circulation to the evaporation vapor on the way. The precipitation in Tibetan Plateau is mainly local rain formed by wet convection from the surface evaporation vapor. The rate of the precipitation formed directly from ocean air mass vapor to the total precipitation in Anduo area is small in summer monsoon.