Intelligent agriculture, also known as smart farming, is an emerging field that integrates advanced technologies and data-driven approaches to optimize agricultural practices. By leveraging artificial intel- ligence, the Internet of Things (IoT), and big data analytics, intelligent agriculture aims to enhance crop yield, minimize resource consumption, and improve overall sustainability. In this context, the accurate detection of diseases in plants holds significant importance for ensuring optimal crop health and productivity. This research proposes an intelligent deeplearning model based on PSO for disease detection in intelligent agriculture. PSO and deep learning are combined to automatically search for the optimal architecture. Utilizing CNNs as the base architecture, the model leverages PSO to determine hyperparameters. Extensive experiments perform better than manual design, with high accuracy, precision, and recall. The model demonstrates robustness and generalizability across diverse datasets, promising practical applications in intelligent agriculture. This research automates architecture design, improving crop management and enabling early disease detection.