Rice is one of the most widely consumed principal foods around the world. However, various diseases and nutritional deficiencies significantly prevent rice crop development, leading to defective and reduced yields. The failure to obtain high-quality crops leads to big financial losses and obstacles in food insurance in some countries. Therefore, effective crop monitoring is crucial for detecting early diseases and deficiencies that threaten yield. Traditionally, diagnosing these problems requires much effort and time from specialized expertise, which is neither scalable nor readily accessible to all agricultural areas, especially in large fields. This study presents a new approach for automating the detection of rice leaf diseases using proven, efficient machine-learning techniques. The proposed model consists of three main phases: data preprocessing, data classification, and model evaluation. In the preprocessing phase, techniques such as data augmentation to increase the dataset, random shuffling of data instances, and normalization were applied to improve the learning process. A custom convolutional neural network (CNN) was then developed to identify rice leaf diseases automatically. Finally, various evaluation metrics were used to assess and compare the model’s performance with previous models. The proposed model achieved an accuracy of 97.03%, a precision of 97.06%, a recall of 97.04%, an F1-score of 97.04%, and an area under the curve (AUC) of 1.0, proving its superiority over previous models. These results suggest that the model can assist farmers in early disease detection and management, enabling timely and accurate interventions.