IoT and Machine Learning-Based Smart Watering Model for Water Optimization in Vegetable Gardens

Abstract

Efficient water management has become increasingly important in modern agriculture due to growing increasing demand for finite water supplies and the necessity of promoting sustainable farming practices. Traditional time-based irrigation approaches often result in inefficient water use and limited adaptability to dynamic environmental conditions. This study presents the design and preliminary validation of an Internet of Things (IoT)- and Machine Learning (ML)-based smart irrigation framework at Technology Readiness Level (TRL) 3. The proposed framework integrates real-time sensor measurements, external weather information, and a Random Forest–based forecasting algorithm to determine crop water demand support adaptive irrigation scheduling. Experimental and simulation-based evaluations demonstrated that the Random Forest model achieved satisfactory predictive performance, with an RMSE of 0.19 L/m² and an MAE of 0.16 L/m². Furthermore, the proposed framework showed the potential to reduce irrigation water consumption by approximately 30% compared with conventional fixed-schedule irrigation while maintaining adequate water availability for crop growth. The integration of multi-source environmental data and predictive analytics enabled more accurate irrigation decisions, contributing to improved water-use efficiency and reduced irrigation-related operational costs. These findings highlight integrating connected sensing systems with machine learning techniques can facilitate evidence-based irrigation management while promoting long-term agricultural sustainability.