Rice, an important food crop, is often faced with widespread and highly contagious diseases that have posed significant threats to agricultural development. However, traditional rice disease identification methods, which rely on the subjective visual examination of experts, hinder timely disease prevention. The advancements in computer vision and deep learning techniques have been key in improving crop disease diagnosis and recognition. This paper suggests a new approach for classifying rice leaf diseases, first via enhancing the input rice plant image through the conversion and extraction techniques. Secondly, by employing the multiple Vision Transformers (ViTs) method, which adopts and dedicates one ViT per type of disease, thereby improving the accuracy and robustness of the disease’s classification in rice plants. The implementation and evaluation of the proposed methodology on 5932 rice plant images demonstrate superior performance in identifying four common kinds of rice leaf diseases, namely bacterial blight, blast, brown spot, and tungro. The precision, recall, and F1-score are significantly improved to become 98%, 98.25%, and 98.125%, respectively. The findings of this work enrich the domain of agricultural technology by introducing an innovative approach to enhance disease management in rice cultivation, which aims to foster robust crop yields to mitigate losses associated with plant diseases.