Smart agriculture incorporates cutting-edge technologies like cloud computing, artificial intelligence, and the Internet of Things (IoT). These networked systems are, nevertheless, more susceptible to cybersecurity risks, especially Distributed De-nial of Service (DDoS) assaults, which have the potential to interfere with vital functions. A machine learning based method for identifying DDoS assaults in IoT-enabled agricultural networks is presented in this paper. XGBoost, Random Forest, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and a CNN-LSTM hybrid model were among the models created and assessed. With a 99.28% accuracy rate, the CNN-LSTM model outperformed the others by successfully identifying both temporal and spatial traffic patterns. Other models also showed strong performance, with Random Forest achieving 97.73%, XGBoost 97.28%, and LSTM 97.11%. The findings demonstrate how well ensemble and deep learning methods can recognize intricate attack patterns in agricultural systems. In order to improve resilience and protect data integrity in smart agricultural contexts, this study highlights the significance of intelligent, adaptive security systems.