Meteorological data is now used widely in modern industrial controlling and energy area. These data concludes not only real-time data, but also includes prediction meteorology data. In this paper, a new kind of intelligent weather station forecasting system based on OMAP-L138 ARM and DSP is presented. This system uses ALN algorithm to realize its prediction function. Through this system, prediction models can be trained by historical data of more than three months, and forecasting results can be obtained by entering real-time meteorological data into the trained models.