Accurate diagnosis of rice pests and diseases is essential but often challenging using traditional methods, which are time-consuming and prone to human error. In this study, we propose the Rice Deep Knowledge Graph (RiceDKG) Expert System, which integrates deep learning techniques, particularly Long Short Term Memory (LSTM), with a Knowledge Graph to enhance symptom pattern-based diagnosis accuracy. This hybrid approach captures relationships among rice plant symptoms while leveraging systematically constructed domain knowledge. The system was evaluated on a dataset of 25 test cases, encompassing various symptoms such as brown spots, leaf curling, and fungal damage. Evaluation results demonstrate an overall accuracy of 84%, with 21 out of 25 cases correctly diagnosed, compared to expert evaluations. These findings indicate that integrating LSTM with knowledge graphs improves the system's ability to handle diverse diagnostic scenarios.