The world is experiencing a rising water crisis with agriculture taking over and using over 70 percent of the freshwater at their disposal. This has been worsened by poor traditional irrigation methods and illegal water diversion through canals and distribution channels resulting in disproportional water distribution and low production of crops. It is not feasible to manually keep track of large irrigation networks, particularly at night, to detect water theft in real-time. To overcome these shortcomings, this paper introduces an loT-based smart irrigation monitoring and water theft detection system. The proposed system installs several sensors of water levels at predetermined points on the irrigation pipeline to continuously check and compare water levels in real time. The sensor data are sent to an loT-enabled communication framework to a cloud interface where they are analyzed and visualized to be remotely monitored through mobile or computer-based applications. The experimental findings through repeated tests indicate that under normal conditions there is a gradual change in water levels, but large drops at downstream areas indicate clearly potential cases of water theft or leakage. The system will be able to detect the specific location and time of such anomalies and produce alerts to intervene in time. The proposed solution will enhance the efficiency of irrigation and make agricultural activities in water-limited areas sustainable by decreasing human input and maintaining equal water distribution and minimizing water wastage.