The problem of disease in rice plants is an obstacle faced by farmers after planting. One of the rice diseases that really worries farmers is disease brown spot, Bacterial Blight, and blast, which causes the leaves to turn yellow prematurely, spots and rice stalks rot. One farmer in Parigi Village whose rice field was attacked by disease suffered a loss of Rp. 6,000,000 to Rp. 8,000,000 per hectare. From all the rice fields of Parigi Village residents. This research aims to detect types of disease in rice plants by applying methods deep learning using YOLO v5 (You only Look Once). The trained model is able to recognize Brown Spot, Bacterial Blight and Blast diseases with a high level of accuracy. In this analysis, two epochs stand out as the best candidates, namely epoch 250 and epoch 200. At epoch 250, the model shows the highest precision (0.802) and a strong mAP@0.5 value (0.702), indicating excellent model performance without overfitting. Meanwhile, at epoch 200, although precision and recall were slightly lower, the highest mAP@0.5:0.95 value (0.393) indicated better generalization ability. Based on these metrics, epoch 150 is identified as the optimal epoch, although epoch 200 also shows strong performance, especially in generalization over a wide range of threshold IoU. The calculation results show the following performance metrics: Precision: 92.5%; Recall: 90.8%; F1-Score: 91.6%; Mean Average Precision (mAP): 93.2%.