Utilizing the real - time water vapor data obtained from GPS water vapor monitoring net in the trial zone and adopting the development of the module and system integration way, the GPS precipitation forecast system for river catchment has been developed. The structure, function, running, and key technology of the system are introduced respectively. Every method of precipitation forecast is simply depicted. An ANN model joined with GA for river catchment precipitation forecast has been set up. The result shows that the GPS water vapor and the model of GA - BP can enhance the forecast precision, and its effectiveness and the reliability has been proved. The system can realize the information processing of different data sources and the precipitation forecast of the result is made by the GPS observed data, supper - air data, satellite cloud atlas etc. of river catchment. It can service for prewarning prediction and prevent flood decision making. 