<p>This study looks into the implementation of artificial intelligence (AI) techniques, particularly deep learning and traditional machine learning models, for the detection of diseases in an aquaponic system with an emphasis on spinach leaf powdery mildew. Since aquaponics incorporates aquaculture and hydroponics, the cultivation method is bound to have plant health management challenges because of its closed-loop system. The study employed sophisticated AI models such as Convolutional Neural Networks (CNNs), Decision Tree (J48), Support Vector Machines, K-nearest neighbors, LogitBoost, Random Forest, and Naïve Bayes to comparatively diagnose powdery mildew and assess their relative performance. To prepare these models, high-quality images of healthy and diseased spinach leaves were collected. The models were evaluated using performance metrics including True Positive Rate, False Positive Rate, Precision, Recall, F-measure, and Matthews Correlation Coefficient. Among these models, CNN outperformed the rest, achieving a robust True Positive Rate (94.44%) and high MCC (0.8681), demonstrating its superior accuracy and reliability. Other models such as J48 and Random Forest also produced reasonable results and could be relied upon in terms of interpretability and accuracy, highlighting the complementary role of classical ML models in practical scenarios. This research provides positive conclusions that there is a possible way for the use of AI technology in promoting disease control in aquaponics, enabling early disease detection, supporting prevention strategies, and guiding resource allocation for sustainable farming practices in difficult climatic regions and water-scarce areas.</p>