Today, agriculture areas in Saudi Arabia, where the water level decreasing resistance in farms growing weeds. Currently the farmers using manually watering for all kind of weeds even if there is no need of water. This is a slow and laborious process when done manually which has led to automated precision weed control as a viable option for such weed watering. Therefore, automatic system is require to decide the present of different weeds. Automatic weed classifi- cation required the ability to identify the accurate class of weed for watering in the given field. That identify the exactly need of a region, this article aims to describe the development of real time automatic water sprayer system for selective weed. In order to improve the agricultural sprayer process, and to make it possible for one sprayer to cover the whole area automatically, the control system for Automatic agricultural irrigation system can recognize using wavelet transform. The system can detect areas where there is no need, little or more water required through the connected camera with a high accuracy. A MATLAB program is developed as software based interface between the system and camera. The software based results will be implement on hardware for real time weed classification.