This paper presents a new localization method for unknown sensor nodes in WSN-based smart irrigation systems, aiming to improve water conservation. The method uses estimated range measurements and the Levenberg–Marquardt (LM) optimization algorithm to accurately determine the positions of unknown nodes, even when they are far from anchors. The proposed method transforms the localization problem into a nonlinear least-squares optimization problem, leveraging known positions of a subset of nodes and inexact distance measurements between pairs of nodes. The method is validated through extensive simulations and experiments, demonstrating significant improvements in estimation accuracy compared to existing algorithms such as DV-Hop and SDR-LM. The results show a 19% and 58% improvement in estimation accuracy, respectively. The method is also robust against measurement noise and scalable for large-scale networks. Integration of this method into smart irrigation systems can reduce water consumption by approximately 28%. The paper also discusses the importance of accurate node localization in smart irrigation systems for efficient water usage and targeted crop management. The proposed method is effective in various network scales and is robust to changes in anchor node locations. The method outperforms other localization algorithms in terms of accuracy and efficiency, leading to significant water savings in smart irrigation systems.This paper presents a new localization method for unknown sensor nodes in WSN-based smart irrigation systems, aiming to improve water conservation. The method uses estimated range measurements and the Levenberg–Marquardt (LM) optimization algorithm to accurately determine the positions of unknown nodes, even when they are far from anchors. The proposed method transforms the localization problem into a nonlinear least-squares optimization problem, leveraging known positions of a subset of nodes and inexact distance measurements between pairs of nodes. The method is validated through extensive simulations and experiments, demonstrating significant improvements in estimation accuracy compared to existing algorithms such as DV-Hop and SDR-LM. The results show a 19% and 58% improvement in estimation accuracy, respectively. The method is also robust against measurement noise and scalable for large-scale networks. Integration of this method into smart irrigation systems can reduce water consumption by approximately 28%. The paper also discusses the importance of accurate node localization in smart irrigation systems for efficient water usage and targeted crop management. The proposed method is effective in various network scales and is robust to changes in anchor node locations. The method outperforms other localization algorithms in terms of accuracy and efficiency, leading to significant water savings in smart irrigation systems.